Anexo 4/Appendix 4 - Presentaciones - Presentations

Anexo 4.1/Appendix 4.1 - Las mayores crecidas en la cuenca del Plata
Referencias bibliográficas
Anexo 4.2/Appendix 4.2 - Controle de cheias nos reservatóriso do setor elétrico brasileiro na bacia do alto paraná
Referências bibliográficas
Anexo 4.3/Appendix 4.3 - The Mississippi basin flood of 1993
References
Anexo 4.4/Appendix 4.4 - Programa de mitigación de inundaciones de la cuenca del río Magdalena-Colombia

Anexo 4.1/Appendix 4.1 - Las mayores crecidas en la cuenca del Plata

por Víctor Pochat

RESUMEN EJECUTIVO

El sistema fluvial del Río de la Plata, con una superficie de cerca de 3.100.000 km², se extiende por los territorios de la Argentina, el Brasil, Bolivia, el Paraguay y el Uruguay, estando formado por tres grandes unidades hidrográficas, el Paraná, el Uruguay y el Paraguay.

Presenta una gran variedad climática, a la cual se agrega la fertilidad de sus suelos, sus reservas minerales, un gran potencial hidroeléctrico y la disponibilidad de vías navegables, lo que ha motivado que se congregue una vasta porción de la población y de la actividad económica de los países que integran su territorio.

Dada su ubicación y extensión, se destaca por su complejidad hidrológica, motivada por notables variaciones en el régimen de precipitaciones y la marcada diversidad geomorfológica, topográfica y de cobertura vegetal de las distintas subcuencas.

La principal fuente de alimentación de los ríos de la Cuenca es pluvial. La precipitación media anual disminuye desde el rango de 2.400-1.400 mm en el este hasta el de 1.000-400 mm en el oeste.

En el norte, el 80% del total anual se concentra en los cinco meses que van desde noviembre hasta marzo, pero este grado de estacionalidad se reduce hacia el sur. En la cuenca del Uruguay casi no hay estacionalidad, siendo la precipitación promedio relativamente uniforme a lo largo del año.

Los caudales medios anuales del río Paraná son de 12.480 m³/s en Posadas y de 16.320 m³/s en Corrientes. En los ríos Paraguay y Uruguay, los valores respectivos son de 3.840 m³/s en Puerto Bermejo y de 4.640 m³/s en Concordia.

La complejidad hidrológica se manifiesta, entre otros aspectos, en la distribución estacional de los caudales anuales. Mientras en Posadas, sobre el río Paraná, esa distribución responde al patrón de precipitaciones, en Puerto Bermejo, sobre el Paraguay, el esquema se revierte totalmente, por efecto del almacenamiento en la zona de expansión denominada Pantanal.

Desde fines del siglo pasado se han registrado crecidas muy importantes, con caudales muy elevados y, en algunos casos, una larga duración.

La mayor parte de esos episodios extremos ocurrieron a partir de mediados de la década del 60, llamando la atención no sólo la magnitud de sus picos (los mayores, 60.200 m³/s en julio de 1983 y 54.000 m³/s en junio de 1992, en Corrientes) - comparable al de las crecidas del siglo XIX y comienzos de este siglo -, sino el aumento de su frecuencia, consecuencia probablemente de los cambios en la incidencia de los sistemas climáticos de gran escala que provocan inundaciones, con implicancias fundamentales en la estimación de los riesgos. En cuanto a la duración, es notable la permanencia del suceso de 1982/3, en el cual los caudales superaron los 30.000 m³/s durante 273 días, en la estación mencionada.

Al tener en cuenta esa realidad, se aconseja basar los estudios estadísticos en muestras posteriores a 1960 y considerar a las crecidas como eventos completos, dando importancia tanto a sus picos como a su duración.

El impacto sobre la actividad social y económica de la región puede evaluarse a partir de las estimaciones de los daños directos e intangibles, los cuales, para las crecidas de 1982/3 y 1991/2 indican valores de 2614 y 513 millones de dólares para los primeros, mientras que para los segundos, cabe mencionar, por ejemplo, 2.350.000 y 3.126.000 ha inundadas, y 234.685 y 122.503 personas evacuadas, respectivamente.

En Argentina, hay actualmente una serie de medidas en curso y previstas para rehabilitar la infraestructura dañada, proteger la vida y propiedad de una gran parte de los pobladores potencialmente afectados, mejorar su calidad de vida, garantizar el normal y pleno desarrollo de las actividades productivas y la integridad, seguridad y funcionamiento ininterrumpido de las vías de comunicación y transporte en el área de influencia de los ríos principales de la Cuenca.

El informe presentado a continuación está dividido en cinco partes:

1. Características Generales
2. Características Hidrológicas
3. Crecidas más Importantes
4. Estimación de Daños por Crecidas
5. Programas de Mitigación

1. Características Generales

El sistema fluvial del Río de la Plata, uno de los mayores del mundo, abarca una superficie de cerca de 3.100.000 km², que se extiende por los territorios de la Argentina, el Brasil, Bolivia, el Paraguay y el Uruguay (Fig. 1). Las coordenadas extremas se ubican entre los meridianos 67°00' y 43°00', y los paralelos 14° 05' y 37° 37' (Ref. 7).

Dentro del sistema cabe diferenciar tres grandes unidades hidrográficas, el Paraguay, el Paraná y el Uruguay.

La cuenca del río Paraguay abarca 1.095.000 km², recorriendo 2.550 km. desde la Chapada de Pareéis hasta su desembocadura en el Paraná. La mayor parte se extiende por una inmensa llanura de naturaleza aluvial, de muy escasa pendiente y con extensas planicies de inundación.

En la parte superior de la cuenca, las márgenes del Paraguay son bajas e inundables y presentan una zona de expansión denominada Pantanal, que cubre cerca de 60.000 km². Este enorme almacenamiento natural tiene una influencia decisiva en el régimen de escurrimiento y en la hidrología de las inundaciones.

El valor medio estimado de la pendiente del río es de cerca de 0,037 m/km.

En el tramo que se extiende desde la desembocadura del río Apa - límite entre el Brasil y el Paraguay - hasta unos 40 km aguas abajo de Asunción, la zona inundable se restringe a un ancho variable de 5 a 10 km y ocupa solamente la margen derecha. El límite inferior de la zona está bien marcado por una quebradura acentuada del perfil longitudinal. Desde este límite hasta el río Paraná, se extiende la zona de desembocadura. En las grandes crecidas el río se desborda sobre las dos márgenes, ocupando una faja de un ancho que varía entre 10 y 15 km.

El Paraná es el río más importante del sistema hidrográfico del Río de la Plata, debido a la magnitud de sus derrames, la extensión de su área tributaria y la longitud de su curso. La superficie de su cuenca abarca 1.510.000 km², sin contar la de la subcuenca del río Paraguay.

Todo el Paraná superior se encuentra en el Brasil, para ser luego y sucesivamente límite brasileño-paraguayo hasta la confluencia con el río Iguazú, argentino-paraguayo y, finalmente, después de confluir con el río Paraguay, internarse en territorio argentino hasta su desagüe en el Río de la Plata.

Su longitud es de 2.570 km desde la confluencia entre el Paranaíba y el Grande - que le dan origen - hasta su desembocadura, y de 3.740 km, si se le agrega el recorrido del primero.

El lecho de su curso superior tiene un perfil escalonado con dos notorios quiebres de la pendiente en Urubupungá y Guaira. Otras caídas menores y la presencia de rápidos caracterizan su perfil longitudinal, consecuente con la estructura geológica y sus accidentes tectónicos.

Figura 1 - Cuenca del Plata

El lecho es de dimensiones muy irregulares: de 4.000 m aguas arriba de Guaira a únicamente 60 m de ancho en Itaipú. Hasta confluir con el Paraguay - cuyas aguas remansa - su anchura fluctúa entre los 150 y los 2.500 m frente a Posadas. Al oeste de esta ciudad el río Paraná diversificaba su cauce presentando una serie de islas, que determinaban la formación de brazos que dividían el escurrimiento fluvial, situación que ha sufrido modificaciones por la presencia del embalse de Yacyretá.

Luego de unirse al Paraguay, el río desciende con muy leve pendiente, 0,036 m/km, hasta su desembocadura. Su ancho va disminuyendo desde los 4.200 m frente a Corrientes a 2.000 m en Rosario. El lecho de inundación es, sin embargo, bastante mayor y extendido casi totalmente sobre la margen derecha, mucho más baja. Los anchos de esta planicie de inundación, que varían entre 13 km frente a Corrientes y 56 km en la sección Rosario-Victoria, señalan la enorme extensión que puede abarcar el Paraná en sus grandes crecidas.

En este tramo medio e inferior, el cauce principal presenta numerosas islas, diseminadas a lo largo de su curso, las cuales son cubiertas totalmente por las aguas de crecida.

Su delta terminal, con numerosos brazos, se inicia a 320 km de la desembocadura y con un ancho que varía entre los 18 y 61 km, cubre un área de 14.100 km². El notable avance frontal de este delta, puede explicarse por la extraordinaria cantidad de sedimentos que transporta el río.

El río Uruguay es el segundo sistema fluvial en importancia, con un desarrollo de cerca de 1.600 km. Nace también en territorio brasileño, aproximadamente a 1.800 m de altura.

Algunos kilómetros aguas abajo de la confluencia con el Pepiri-Guazú, donde el río Uruguay pasa a ser límite entre la Argentina y el Brasil, se estrecha entre paredes rocosas, formando cataratas donde el río pasa de 450 a cerca de 20 m de ancho.

La cuenca superior se desenvuelve sobre rocas de origen volcánico. En la parte centro-oeste de la región aparecen otros tipos litológicos entre los que se destaca el aluvión a lo largo de los ríos Santa María e Ibicuí.

Sus afluentes más importantes provienen del este. Son ríos con cuencas relativamente pequeñas, pero de considerable caudal. El Quaraí sirve de límite entre el Brasil y el Uruguay. El mayor tributario del Uruguay es el río Negro, que desemboca a menos de 100 km del estuario del Plata. Luego de esa confluencia, el Uruguay se ensancha convirtiéndose, virtualmente, en una extensión del Río de la Plata, cuya cuenca propia abarca otros 130.000 km².

Dada su gran extensión, la cuenca del Plata presenta una gran variedad de condiciones climáticas: tropical en la cabecera del río Paraguay; subtropical en las llanuras del Chaco argentino-paraguayo, en el nordeste argentino y en los estados del sudeste brasileño; templado en el sur de la Mesopotamia Argentina - entre los ríos Paraná y Uruguay - en el sur de Brasil y en el Uruguay; seco en las nacientes de los tributarios en Bolivia y en el noroeste de Argentina.

Esta variedad climática, a la cual se agrega la fertilidad de sus suelos, sus reservas minerales, el potencial hidroeléctrico de sus grandes ríos y la disponibilidad de vías navegables, ha congregado una vasta porción de la población y de la actividad económica de los países que integran el territorio de la Cuenca. (Ref. 3)

Figura 2 - Marcha anual de los caudales medios mensuales en porcentaje del caudal medio anual

Rio Paraná en posada 1901-1992

Rio Paraguay en puerto Bermejo 1910-1993

Rio Paraná en Corrientes 1904-1993

Rio Uruguay en concordia 1898-1993

2. Características Hidrológicas

La principal fuente de alimentación de los ríos de la Cuenca es pluvial; sólo nieva en las cabeceras de algunos de ellos, representando un porcentaje muy bajo del total. La precipitación media anual disminuye desde el rango de 2.400-1.400 mm en el este hasta el de 1.000-400 mm en el oeste (Ref. 7). En las cuencas del Iguazú y del Alto Uruguay, las lluvias anuales medias son las más altas de la Cuenca del Plata.

En el norte, el 80% del total anual se concentra en los cinco meses que van desde noviembre hasta marzo, pero este grado de estacionalidad se reduce hacia el sur. En la cuenca del Uruguay casi no hay estacionalidad, siendo la precipitación promedio relativamente uniforme a lo largo del año (Ref.5).

En cuanto a los caudales, se dispone de registros sistemáticos a partir de principios de este siglo. El Cuadro 1 presenta los caudales medios anuales en ciudades de Argentina.

Cuadro 1 CAUDAL MEDIO ANUAL EN DETERMINADAS CIUDADES DE LA CUENCA DE LOS RÍOS PARANÁ, PARAGUAY Y URUGUAY

La complejidad hidrológica de la Cuenca del Plata comienza a surgir cuando se examina la distribución estacional de los caudales anuales (Fig.2). En Posadas, sobre el río Paraná, esa distribución es la esperada en base al patrón estacional de precipitaciones, teniendo en cuenta los efectos de retardo en una cuenca tan grande; los caudales promedio más altos ocurren entre diciembre y abril.

En Puerto Bermejo, sobre el río Paraguay, el esquema se revierte completamente. Los flujos de entrada al Pantanal siguen la distribución estacional esperada, pero su almacenamiento natural modifica el flujo de salida, ocasionando se demore hasta seis meses. Aguas abajo, los caudales promedio más altos ocurren entre agosto y noviembre.

En Corrientes, sobre el río Paraná, la integración de esos dos regímenes de caudal tiene como consecuencia una distribución más pareja del caudal promedio a lo largo del año. Podría suponerse que la compensación de las estaciones de caudales pico en el Paraná y el Paraguay reduce la intensidad del riesgo de inundación aguas abajo de la confluencia. Sin embargo, esto no es enteramente así, ya que la distribución estacional del riesgo de inundación es más compleja que lo que sugiere la marcha anual de los caudales medios mensuales (Ref.5).

En Concordia, sobre el río Uruguay, dicha marcha no parece reflejar, a primera vista, la distribución estacional de precipitaciones para la cuenca superior, que es relativamente uniforme mes a mes, mientras que los caudales se presentan con un período intenso de caudales bajos entre enero y marzo. Esta discrepancia parecería sugerir que en dichos meses el escurrimiento es bajo. Sin embargo, ésta es una conjetura que revela lo poco que se comprende el funcionamiento espacial y temporal de lluvias y su relación con la intensidad y duración de las crecidas.

3. Crecidas Más Importantes

Desde fines del siglo pasado ocurrieron una cantidad de eventos importantes. En el Cuadro 2 se muestran los diez episodios más extremos, en cuanto a caudal pico, en cuatro sitios. De esta información, surge que sobre el río Paraná y el río Paraguay, la mayor parte de los eventos más importantes ocurrieron a partir de 1965 y en cuanto al río Uruguay, cuatro de los eventos más importantes desde 1898 ocurrieron a partir de 1972.

Cuadro 2 - ORDENAMIENTO DE LOS CAUDALES PICO EN CUATRO SITIOS EN LA CUENCA DEL PLATA

Río Paraná en Posadas 1901/2-1991/2

Río Uruguay en Concordia 1898 - 1993

Si el mecanismo de generación de excedentes hubiera sido homogéneo a lo largo del tiempo, estos datos serían estadísticamente improbables. Al no haber sido así, significa que la incidencia de las crecientes en la Cuenca del Plata aumentó en las últimas dos o tres décadas (Ref.5).

Existen referencias sobre crecidas extraordinarias ocurridas en Corrientes en 1612 y 1748, mientras que la disponibilidad de datos en cuanto a niveles fluviales del siglo XIX permite realizar estimaciones del caudal asociado a las crecidas notables del 1812, 1858 y 1878. (Cuadro 3).

Cuadro 3 RIO PARANÁ EN CORRIENTES: ORDENAMIENTO DE LOS SEIS CAUDALES EXTREMOS MÁS IMPORTANTES (1800-1993)

Figura 3 - Ríos Paraná y Paraguay, hidrogramas de los cuatro mayores eventos (en términos de pico) registrados en Corrientes, posadas y pto.bermejo (A)

Figura 3 - Ríos Paraná y Paraguay, hidrogramas de los cuatro mayores eventos (en términos de pico) registrados en Corrientes, posadas y pto.bermejo (B)

Figura 3 - Ríos Paraná y Paraguay, hidrogramas de los cuatro mayores eventos (en términos de pico) registrados en Corrientes, posadas y pto.bermejo (C)

Figura 3 - Ríos Paraná y Paraguay, hidrogramas de los cuatro mayores eventos (en términos de pico) registrados en Corrientes, posadas y pto.bermejo (D)

Los valores estimados podrían sugerir que la magnitud de los picos de crecidas recientes no es excepcional, sino que el aumento de la frecuencia de los eventos extremos a partir de 1960 representa un cambio significativo en la incidencia de los sistemas climáticos de gran escala que provocan inundaciones.

En cuanto a la duración de los mayores episodios de inundación de este siglo en la cuenca del río Paraná/Paraguay, es notable la permanencia del suceso de 1982/3 (Figura 3 y Cuadro 4).

Cuadro 4 RÍO PARANÁ EN CORRIENTES: DURACIÓN DE DIVERSAS CRECIDAS REGISTRADAS

La explicación para la duración del episodio de 1982/3 es la secuencia de lluvias mensuales registradas en un solo período de doce meses. Entre junio de 1982 y mayo de 1983, las precipitaciones de la cuenca en junio, noviembre, enero y mayo fueron las mayores jamás registradas en esos meses desde comienzos del siglo. Por lo tanto, el evento resultante debe considerarse sumamente raro teniendo en cuenta la distribución probabilística conjunta de la magnitud de la lluvia, su duración, y por la poca probabilidad de que cuatro lluvias máximas mensuales ocurran en un mismo año.

Los eventos 1905/06, 1989/90 y 1991/2 representan episodios de inundación más típicos sobre el río Paraná y tienen mucho en común; en especial las inundaciones de 1905/6 y 1991/2. Una característica particular - y llamativa para una cuenca tan grande - es la rapidez con que el hidrograma crece en cada uno de estos eventos. En las dos semanas anteriores al arribo del pico a Comentes, la tasa de crecimiento de los tres eventos es casi idéntica (1.400 m³/s por día). Los hidrogramas son notablemente simples en cuanto a la forma; sólo el correspondiente al episodio de 1992 presenta aportes diferenciables de distintas subcuencas de aguas arriba. En Corrientes, las tasas de recesión también son rápidas, y por lo general constituyen alrededor del 50% de la tasa creciente (700 m³/s por día) (Fig. 3).

Estos datos indican que la respuesta de la cuenca del Paraná aguas arriba de Corrientes, ante lluvias breves (semanas más que meses), es relativamente uniforme. En cambio, las crecidas muy severas se producen por eventos de lluvia en escala sinóptica, por lo que grandes áreas se ven afectadas por precipitaciones inusualmente prolongadas y extensas.

Figura 4 - Rio Uruguay, hidrogramas de los cuatro mayores eventos (en términos de pico) registrados en concordia (A)

Figura 4 - Rio Uruguay, hidrogramas de los cuatro mayores eventos (en términos de pico) registrados en concordia (B)

Figura 4 - Rio Uruguay, hidrogramas de los cuatro mayores eventos (en términos de pico) registrados en concordia (C)

Figura 4 - Rio Uruguay, hidrogramas de los cuatro mayores eventos (en términos de pico) registrados en concordia (D)

Las condiciones más críticas para producir crecidas en Posadas y Corrientes se asocian con lluvias en toda la cuenca del Paraná, provocando caudales de base altos y caudales pico de la parte superior de la cuenca, con posteriores tormentas concentradas en la parte inferior de la cuenca superior, como ocurrió en 1982/3 y 1992 (Ref. 2).

Tal como se ve en la Figura 3, el aporte de los caudales del río Paraguay al pico de caudal en Comentes no es relevante frente a las descargas que llegan desde Posadas. Las inundaciones en la parte inferior de la cuenca del Paraguay, aguas abajo de Asunción, no se relacionan tanto con el caudal en el propio río, sino con el efecto de remanso aguas arriba de la confluencia con el Paraná.

En la Figura 4, se presentan los hidrogramas de los cuatro eventos de crecida del río Uruguay en Concordia, más grandes por su pico, desde 1898. El caudal es mucho más variable que sobre el Paraná, como es de esperarse en una vertiente mucho menor, cuyo almacenamiento natural es tan distinto. La distribución temporal de las lluvias se ve más claramente en la variación de caudales a lo largo del año, y resulta evidente el aporte de los distintos afluentes y subcuencas. La falta de almacenamiento queda manifiesta con los tres episodios de 1983, separados por períodos de caudales mucho menores. Este hidrograma puede compararse con el de Corrientes del mismo año, donde el impacto del almacenamiento aseguró la continuidad de caudales elevados por más tiempo. Aun así, el evento de 1983 sobre el Uruguay, al igual que sobre el Paraná, sigue siendo el de más larga duración de caudales extremos, según indica el Cuadro 5.

Cuadro 5 Río Uruguay en Concordia: Duración de diversas crecidas registradas

En el Cuadro 6 se muestra el tiempo y la velocidad de traslado de los caudales pico en el río Paraná.

Cuadro 6 Río Paraná: Tiempos medios de tránsito de los caudales y velocidades en determinados tramos

Para el río Paraguay, la incidencia estacional de las inundaciones concuerda con la distribución de los caudales medios mensuales, aunque podría haberse esperado que la frecuencia de los picos de caudal en Puerto Bermejo fuera mayor más avanzado el año. El hecho de que estos picos se limiten al período entre mayo y julio refleja el impacto del almacenamiento en el Pantanal y la atenuación relativamente constante de los caudales de crecida.

En cuanto a la estacionalidad, el régimen de crecida más complejo es el del Uruguay, donde se dan dos estaciones de crecida bien diferenciadas, y donde la mayor probabilidad de que el pico de caudal máximo anual ocurra es en octubre (Ref. 5). El Uruguay y el Paraná están, por lo general, desfasados en cuanto a la ocurrencia de crecidas en un año determinado, hecho que se relaciona, sin duda, con el tamaño de la cuenca, la diferencia de almacenamiento natural y el régimen pluvial.

Sin embargo, este panorama de la distribución estacional de los caudales pico anuales resulta inconsistente con la ocurrencia de los mayores eventos registrados. De las diez mayores crecidas registradas en Posadas desde 1900, seis ocurrieron entre mayo y julio (ver Cuadro 2), cuando la probabilidad promedio de que haya aguas altas es relativamente baja. Esto se entiende por el hecho de que los eventos más extremos son la consecuencia de estados de caudal iniciales muy altos que deberían ocurrir en febrero y marzo. Además, se agregan los volúmenes provenientes de tormentas tardías, que tienden a generar los estados más críticos, tal como ocurrió en 1904/5, 1982/3, 1991/2, 1986/7, 1922/3 y 1928/9.

Sobre el Uruguay, existe una discrepancia similar entre la ocurrencia, en un año, de los estados de crecida "promedio" y extremas. Sólo uno de los diez eventos más importantes ocurrió en octubre, y siete tuvieron el pico entre abril y julio, aparentemente la menor de las dos estaciones de caudales altos.

Cualquier evento de crecida mayor en una cuenca significa apartarse del estado promedio, pero la ocurrencia y la duración de dicho apartamiento, por lo general, sigue un modelo, que a pesar de ser variable, es razonablemente bien definible. En base a este análisis de la historia y las características de la incidencia y la severidad - dada por la magnitud de los picos - de las crecidas, es evidente que éste no es el caso de la Cuenca del Plata, ya que su hidrología de las crecidas es muy compleja, tanto en el espacio como en el tiempo.

El cambio en el régimen hidrológico regional se puede ver claramente graneando los caudales medios mensuales de las cuencas principales antes y después de 1960 (Figura 5). Los resultados son altamente significativos, tanto en el contexto estadístico como hidrológico. Hay una gran consistencia en distintos puntos dentro de la extensa cuenca, lo cual sólo puede justificarse por un cambio sistemático del régimen pluvial regional, tanto en cantidad total como en el patrón de estacionalidad.

Desde el punto de vista de las crecidas y de su ocurrencia, una consecuencia fundamental es que, en promedio, los caudales extremos partirán de caudales iniciales más altos que antes. Por lo tanto, se espera que aumenten los picos de caudal y la duración de los eventos (Ref. 5).

En un estudio del Banco Mundial (Ref. 2) se analizan tres cambios importantes que ocurrieron a partir de la década del '60, los que podrían haber afectado la hidrometeorología de la cuenca: cambios en el uso de la tierra, en especial la desforestación, que pudieron haber aumentado el escurrimiento; el desarrollo hidroeléctrico, principalmente en el Alto Paraná en Brasil, que provocó una mayor regulación y un régimen de caudales bajos más alto, y las precipitaciones durante la estación húmeda, que han sido más altas por lo menos desde 1960, y hacen suponer cambios climáticos, que pueden estar inducidos por el hombre.

Luego de un análisis sistemático de los datos sobre caudal (principalmente en Posadas y Corrientes), sus autores han concluido que las inundaciones fueron más frecuentes y más severas en la segunda mitad de este siglo; que, en el mismo período, los caudales bajos extremos han sido menos frecuentes y menos agudos; que el derrame anual total aumentó y la distribución estacional cambió; que las variaciones pluviales son las mayores responsables de los cambios en los caudales y en las inundaciones; que los cambios en la distribución estacional de los caudales también parecen estar relacionados con los cambios en el régimen pluvial, además del impacto de la operación de los embalses de aguas arriba y, finalmente, que no se encontró ninguna evidencia consistente, ni estadística ni de otro tipo, de que los cambios en el uso de la tierra hayan influido significativamente en el aumento de la incidencia y la severidad de las crecidas recientes.

También se suele considerar que el fenómeno de El Niño de 1982/3 fue el más importante del siglo. En el estudio citado (Ref. 2) se indica que cuando la corriente de El Niño es anómala, las aguas cálidas ecuatoriales del Pacífico se encuentran mucho más hacia el este de lo usual, provocando el típico modelo de intensas lluvias sobre la Cuenca del Plata y a lo largo de la costa norte de Perú, haciendo que el clima se torne más seco en la región noreste de Brasil. En la última década ha habido eventos de El Niño moderados e importantes en 1982/3, 1986/7 y 1991/2, correspondientes a la mayoría de los episodios de crecida en los ríos Paraná, Paraguay y Uruguay. Se puede concluir que es conveniente conocer las cuestiones climáticas asociadas a dichos episodios, a fin de poder predecir crecidas, si bien se debe anotar que las características espaciales y temporales de las anomalías climáticas no son del todo coherentes ni se han explicado suficientemente.

Figura 5 - Caudales medios mensuales, serie completa y serie 1960-1992

Rio Paraná en posadas

Rio Paraguay en Puerto Bermejo

Rio Paraná y Paraguay en Corrientes

Rio Uruguay en concordia

Ya se ha visto que, en general, ha aumentado la incidencia de las crecidas; el paso siguiente será establecer si la severidad cambió a través del tiempo. Existen evidencias de cambios en la severidad de los picos de crecida sobre el Paraná, pero se limitan a las cuencas media e inferior. En Jupiá y Guaira, las cantidades posteriores a 1960 caen dentro del intervalo de confianza del 90% de las cantidades estimadas de la muestra total. Esto no sucede ni en Posadas ni en Comentes, ni sobre el río Uruguay (Ref. 5).

Por lo expuesto, se concluye que el régimen de crecidas de la Cuenca del Plata ha cambiado en forma significativa por lo menos a partir de 1960, tanto en cuanto a la incidencia como a la severidad de las inundaciones. La magnitud de dichos cambios es tal que sólo pueden explicarse por un cambio sistemático a nivel de la circulación atmosférica, lo que influye en el régimen pluvial.

El hecho de que existan evidencias estadísticas e hidrológicas fuertes sobre la discontinuidad del régimen de crecidas en la Cuenca del Plata tiene consecuencias fundamentales en la estimación del riesgo de inundación.

Los estudios estadísticos recientes (Ref. 5) se basan solamente en las muestras posteriores a 1960, considerando que estos datos representan el régimen hidrológico contemporáneo de las cuencas en estudio. No puede determinarse si los cambios posteriores a 1960 son permanentes o no, pero es conveniente señalar que podrían deberse a cambios inducidos por el hombre. Teniendo en cuenta esta posibilidad, no es aconsejable estimar el riesgo de caudales extremos usando todos los datos de los registros, ya que se subestimaría la magnitud del riesgo de crecida. Tal como se indica en el Cuadro 7, si se asocia un período de retorno al pico del evento de 1982/3 con el registro completo, los errores no serían menores.

Cuadro 7 RÍOS PARANÁ Y URUGUAY. PICOS DE CRECIDA E INTERVALOS DE RETORNO ASOCIADOS, CONSIDERANDO LA POBLACIÓN MUESTRAL COMPLETA Y EL REGISTRO POSTERIOR A 1960

Por otra parte, no es suficiente usar solamente los caudales pico de las crecidas para describir el riesgo de inundación en cuencas muy extensas. La duración del suceso puede ser tan importante como el caudal, en especial en cuanto al daño económico. En el análisis multivariado de la incidencia y de la severidad de la inundación, resulta fundamental distinguir entre el pico de crecida, definido sólo por la crecida máxima, y el evento completo, definido por el hidrograma total, que incluye el pico de crecida más su duración más su volumen.

Al considerar la crecida como un evento, el intervalo de retorno de ese evento puede ser muy diferente al del pico. Dos eventos con los mismos picos de crecida pueden tener distintos intervalos de retomo, si uno de ellos tuvo más duración y por lo tanto más volumen.

El tema clave para encontrar una solución estadística operativa para el análisis de las crecidas como eventos, es que el volumen y la duración están íntimamente relacionados. Esto reduce el problema analítico a la identificación y estimación de un modelo de valores extremos bivariado del caudal pico y el volumen.

El Cuadro 8 compara el intervalo de retomo estimado del pico con el del evento de crecida total para los tres mayores eventos. Estas estimaciones sobre el riesgo integral resultan intuitivamente más interesantes que aquéllas basadas sólo en los picos; en especial, demuestran la combinación de pico y volumen de lo sucedido en 1982/3, que lo hacen tan excepcional en la historia de las crecidas.

Cuadro 8 COMPARACIÓN ENTRE LOS INTERVALOS DE RETORNO DE EVENTOS COMPLETOS Y DE CAUDAL PICO

4. Estimación de los Daños por Crecidas

La información histórica disponible permite estimar los daños directos de las inundaciones de 1982-83 y 1992, aunque con importantes salvedades respecto al grado de error y de comparabilidad entre los valores obtenidos (Ref. 5).

Estas salvedades se refieren a superposiciones en la contabilidad de algunos daños, o bien a daños que habrían sido omitidos por información incompleta; al empleo de metodologías que arrojan errores de cálculo significativos; a cambios importantes en la década del 80 respecto al grado de desarrollo de recintos y defensas para los núcleos urbanos más afectados; a falta de coincidencia, en varios casos, entre los valores aportados, por diversas fuentes, y a ausencia de cierta información.

Organismos oficiales y privados, investigadores, periódicos, etc., provinciales y nacionales, han realizado estudios y evaluaciones valiosas y aportado datos muy importantes.

Esta información y estudios, si bien dispersos y no siempre coincidentes en sus conclusiones, merecerían ser compilados y analizados, a fin de compensar el volumen de información directa sobre daños, que se perdió y que es irrecuperable como tal.

En las ciudades es donde hay que concentrar los esfuerzos, para mayor precisión en las evaluaciones de daños, por ser las áreas donde se concentra la mayor riqueza regional y la mayor parte de la población afectada.

Los valores básicos de los daños por las crecidas de 1982/83 y 1992 son los que figuran en los Cuadros 9 y 10.

Cuadro 9 DAÑOS POR LAS CRECIDAS 1982-83 DE LOS RÍOS PARANÁ, PARAGUAY Y URUGUAY (MILLONES DE DÓLARES DE AGOSTO 1983)

Cuadro 10 - DAÑOS POR LA CRECIDA 1992 DE LOS RÍOS PARANÁ, PARAGUAY Y URUGUAY (En millones de US dólares de diciembre 1992)

La información disponible global sobre la crecida de 1992 es menos completa que en el caso de la crecida 1982/83 y, por lo tanto, los datos sobre daños directos pueden tener un grado de error mayor.

Sena útil contabilizar los daños para las diferentes crecidas a un mismo nivel de la actividad económica (Cuadro 11). Sin embargo cabe aclarar que se trata de una estimación, ya que no se cuenta con información económica suficientemente desagregada, que muestre la evolución de la economía en las áreas realmente afectadas. Sólo es posible disponer de esa información por provincias, a nivel departamental y en algunos casos y períodos.

Cuadro 11 - DAÑOS DIRECTOS PARA DIFERENTES CRECIDAS, PARA EL NIVEL DE ACTIVIDAD ECONÓMICA DEL AÑO DE CADA CRECIDA (en millones de dólares, de diciembre de 1992)

El único ajuste que se puede realizar, con relativa confiabilidad, a los daños históricos, es el referido a la moneda, calculando los daños en moneda constante a una fecha dada (Ref. 5).

El valor de los daños por la crecida de 1992 resulta significativamente bajo, para la segunda crecida en importancia en cuanto a pico, siendo sensiblemente inferior, incluso, al de la crecida de 1966 (856 x 10⁶ US$ de 1992) para un pico menor (Ref. 5).

Se podría explicar en parte este resultado teniendo en cuenta la corta duración de la crecida y la existencia de recintos en la mayoría de los principales centros urbanos más amenazados por crecidas anteriores (Clorinda, Formosa, Goya, Resistencia, Santa Fe, Ibicuy y Paranacito) sobre el río Paraná.

Este segundo factor ha sido sustantivo para reducir los daños en que esos núcleos urbanos, salvo en los casos de Clorinda e Ibicuy, donde las aguas atravesaron o sobrepasaron las defensas urbanas, que no eran consideradas definitivas.

No se ha podido esclarecer si esta cifra baja de daños directos se debe exclusivamente a la presencia de tales recintos, o habría también errores de sub-evaluación de los daños.

Existe también información referida a daños intangibles, en particular respecto a áreas inundadas (Cuadro 12) y población evacuada (Cuadro 13).

Cuadro 12 - AREAS INUNDADAS SEGUN DISTINTAS CRECIDAS

Cuadro 13 POBLACION EVACUADA SEGUN DISTINTAS CRECIDAS

Estas cifras reflejan, aunque de un modo limitado, el verdadero significado de los daños no medibles económicamente, pero que afectan las estructuras sociales y la calidad de vida de los sectores poblacionales más directamente afectados, al incidir en los parámetros de salud, de alimentación, de ocupación, etc.

Cabría agregar consideraciones en relación a daños indirectos como, por ejemplo, los impactos de las pérdidas agrícolas sobre la industria procesadora de alimentos y luego sobre otros sectores, que pueden ser significativos. Se han ideado algunas metodologías para estimarlos, basadas en el cálculo de pérdidas de valores agregados para las distintas actividades económicas (Ref. 5).

5. Programas de Mitigación

Ante la grave situación producida por la crecida de 1992, el Gobierno de la República Argentina solicitó ayuda al Banco Mundial, aprobándose el Programa de Rehabilitación para la Emergencia de las Inundaciones por un monto de u$s 318 millones, con financiamiento parcial del Banco mencionado y aportes del Estado Nacional y de las provincias afectadas (Ref. 4).

El Programa se ejecuta en el ámbito de la Secretaría de Asistencia para la Reforma Económica Provincial del Ministerio del Interior, estando a cargo de la Subunidad Central de Coordinación para la Emergencia (SUCCE) la supervisión y coordinación del accionar de Subunidades Provinciales, organismos encargados de su ejecución.

Su objetivo principal es restaurar la normal actividad económica y social de aproximadamente 4 millones de personas, en la región afectada por las inundaciones y suministrar un marco institucional sólido para la coordinación e implementación del programa global, con la finalidad de rehabilitar la infraestructura dañada por esas inundaciones.

Entre las acciones que abarca, cabe señalar la rehabilitación de obras de infraestructura (esencialmente de transporte, salud, educación y saneamiento) y de obras provisorias de atenuación ante crecidas; la reparación y construcción de viviendas para los pobladores ribereños de bajos recursos; el mejoramiento del sistema de pronóstico y de alerta hidrológico y la definición de los términos de referencia de los estudios básicos para el diseño de una solución integral al problema de las inundaciones en la región.

Como complemento indispensable del Programa de Rehabilitación, se encuentra en preparación un Proyecto de Protección contra las Inundaciones, a ser llevado a cabo entre 1996 y 2001, por un monto de US$ 760 millones, con un financiamiento donde también interviene el Eximbank de Japón.

Tiene como objetivos proteger la vida y propiedad de 5,5 millones de habitantes y mejorar su calidad de vida, y garantizar el normal y pleno desarrollo de las actividades productivas y la integridad, seguridad y funcionamiento ininterrumpido de las vías de comunicación y transporte en el área de influencia de los ríos Paraná, Paraguay y Uruguay.

Como acciones a desarrollar, se prevén, entre otras, reforzar o complementar defensas y obras hidráulicas complementarias en lugares seleccionados; mejorar y ampliar los cauces naturales, para drenajes superficiales de las áreas urbanas protegidas y a proteger; definir pautas de planificación e implementación de las medidas de mitigación y el perfeccionamiento en la recolección y uso de la información, para el proceso de análisis y planificación; regular el uso del suelo para la planificación regional y local; establecer los aspectos institucionales y legales que garanticen el manejo integrado de la llanura de inundación; mejorar las medidas de alerta de crecidas, para hacer frente a futuras inundaciones, en particular, en zonas de menor prioridad económica.

AGRADECIMIENTOS

El presente trabajo ha sido elaborado en base a valiosos estudios sobre la Cuenca del Plata, entre los cuales se destaca el preparado por un equipo integrado por la consultora Sir William Halcrow & Partners Ltd. y profesionales independientes, para la Subunidad Central de Coordinación para la Emergencia de la Secretaría de Asistencia para la Reforma Económica Provincial del Ministerio del Interior de la República Argentina (Ref. 5).

Referencias bibliográficas

1. Aisiks, E.G., La gran crecida del río Paraná de 1983, Revista de la Organización Techint, Buenos Aires, 1984.

2. Anderson, R.J., N. da Franca Ribeiro dos Anjos y H.F. Díaz, An Analysis of Flooding in the Paraná/Paraguay River Basin, The World Bank, Latin America & the Caribbean Technical Department, LATEN Dissemination Note 5, Washington, D.C., 1993.

3. Barberis, J., La Plata River Basin, United Nations Interregional Meeting on River and Lake Basin Development with Emphasis on the Africa Region, Addis Abeba, 1988.

4. Ministerio del Interior, Subunidad Central de Coordinación para la Emergencia, Programa de rehabilitación para la emergencia de las inundaciones. Informe descriptivo, estado de avance y estadísticas al 31/8/95, Buenos Aires, 1995.

5. Ministerio del Interior, Subunidad Central de Coordinación para la Emergencia, Estudio de regulación del valle aluvial de los ríos Paraná, Paraguay y Uruguay para el control de las inundaciones, Buenos Aires, 1994.

6. Ministerio de Obras y Servicios Públicos, Secretaría de Recursos Hídricos, Informe de daños. Crecida 1982-83 de los ríos Paraná, Paraguay y Uruguay, Buenos Aires, 1984.

7. Secretaría General de la Organización de los Estados Americanos, Cuenca del Río de la Plata. Estudio para su planificación y desarrollo. Inventario de datos hidrológicos y climatológicos, Washington, D.C., 1969.

Anexo 4.2/Appendix 4.2 - Controle de cheias nos reservatóriso do setor elétrico brasileiro na bacia do alto paraná

Por Alcides Lyra Lopes e Vinicius Forain Rocha

RESUMO

O objetivo deste trabalho é relatar a experiencia do Setor Elétrico Brasileiro na área de controle de cheias na bacia do Alto Paraná. Sao apresentadas as causas mais relevantes que sensibilizaram esse setor a incluir, no planejamento da Operação energética, medidas para a reduçao dos impactos das inundaçoes nas áreas de influência dos seus reservatórios. É feita uma descriçao sucinta das principais características físicas dos aproveitamentos hidroelétricos e dos problemas inerentes ao conflito entre a geraçao de energía e o controle de cheias.

O trabalho da ênfase as metodologías de cálculo dos volumes de espera e ás regras de operaçao estabelecidas pelo GCOI, assim como relata outras medidas para aumentar a eficiencia do controle de cheias. Sao apresentados alguns resultados práticos positivos, obtidos em eventos de cheias ocorridos em 1992, bem como as perspectivas de aprimoramento e desenvolvimento nesta área.

ABSTRACT

This paper describes the experience of the Brasilian Electric System in flood control at Parana river basin, and present the factors that caused the inclusion of the consideration of flood impacts in the energy operation planning studies, there is a brief description of the main characteristics of hydraulic plants and the problems caused by the conflict between energy generation and flood control.

This paper emphasizes the adopted methodologies to calculate the flood control volumes and operation rules stabilished by GCOI, and states additional ways to increase flood control efficiency. We also present practical results obtained in 1992 and possible improvements in this area.

SUMARIO

1) Introduçao
2) O Problema do Controle de Cheias para o Setor Elétrico
3) Planejamento para Operação de Controle de Cheias
4) Aplicaçao á Bacia do Alto Paraná
5) Análise da Experiencia Operativa

1. Eletrobras Introduçao

O sistema brasileiro de geraçao de energía elétrica é um sistema hidrotérmico com forte predominância de geraçao hidráulica. Em 1994, a geraçao hidráulica foi responsável por mais de 98% da energía produzida de 271.300 GWh. Neste ano, a capacidade total instalada no país, considerando 50% de Itaipu, era de 55.400 MW e as usinas hidroelétricas representavam 93% deste total. As instalaçoes de geraçao e transmissao pertencem a diversas empresas e estao interconectadas segundo dois grandes Sistemas Interligados: Norte/Nordeste e Sul/Sudeste.

Esse parque hidroelétrico é hoje constituido por um conjunto de mais de setenta usinas e reservatórios, localizados em varias bacias hidrográficas. A interdependência operativa entre as usinas e a interligaçao dos sistemas de transmissao das empresas trouxeram a necessidade da coordenaçao da operaçao do sistema. Nos sistemas interligados Norte/Nordeste e Sul/Sudeste esta tarefa cabe ao Grupo Coordenador para Operaçao Interligada - GCOI, criado pela Lei 5899 de 05 de julho de 1973 e regulamentado pelo Decreto 73102 de 07 de novembro de 1973, sendo formado pelas empresas estaduais e federáis de energía elétrica sob a coordenaçao da ELETROBRAS, através da sua Diretoria de Operaçao de Sistemas.

Os aproveitamentos hidráulicos do setor elétrico brasileiro, embora planejados e construidos visando a produçao de energía elétrica, vêm também sendo utilizados para múltiplas finalidades, dentre as quais destaca-se o amortecimento de cheias.

2. O Problema do Controle de Cheias para o Setor Elétrico

O efeito de regularizaçao propiciado pela criaçao dos reservatórios dos aproveitamentos hidroelétricos faz com que ajusante dos mesmos as cheias de menor porte, ou de menor tempo de recorrência, ocorram menos freqüentemente. Em conseqüência deste fato, áreas ribeirinhas a jusante desse aproveitamentos, que eram periódicamente inundadas, passaram a ser mais seguras e, portante, atrativas para utilizaçao e ocupaçao. O aproveitamento econêmico e social dessas áreas pelas populaçoes e instituiçoes locais traz para o Setor Elétrico novas variáveis a serem consideradas na operaçao dos sistemas.

Essas novas variáveis, constituidas pelas descargas de restriçao dos aproveitamentos, criam impactos na forma de operar os reservatórios do Setor Elétrico, pois ao objetivo primario de geraçao de energia soma-se o de controle de cheias.

Sob o ponto de vista da produçao de energia, a ocorrência de precipitaçoes intensas na estaçao chuvosa é sempre desejável, pois permite recuperar os volumes armazenados nos reservatórios e aumentar a capacidade de atendimento ao mercado. Por outro lado, sob o ponto de vista de segurança do aproveitamento e dos agentes econômicos instalados em regioes ajusante dos reservatórios, as chuvas intensas e as elevadas afluências decorrentes podem significar riscos de rompimento das restriçoes e prejuízos ambientáis, sociais e econômicos de vulto.

Para efetuar o controle de cheias, o Setor Elétrico prevé a alocaçao de volumes vazios nos reservatórios, durante os períodos chuvosos das regioes onde estes se localizam, capazes de absorver parcelas determinadas das afluências, evitando ou reduzindo o impacto de danos causados por inundaçoes a jusante. E fácil entender que estas medidas geram um conflito, pois contrariam frontalmente o objetivo da produçao de energia elétrica, cujo interesse é utilizar os reservatórios sempre que possível, em sua capacidade máxima de armazenamento.

A manutençao de volumes vazios nos reservatórios do Sistema Interligado conduz a um aumento do risco de que os mesmos nao sejam totalmente recuperados até o final da estaçao chuvosa. Isto traz como implicaçao urna reduçao nas disponibilidades energéticas e, por se tratar de um sistema hidrotérmico, um aumento do risco de geraçao térmica futura e de déficit de energia.

O estabelecimento dos valores ótimos dos volumes vazios a serem alocados nos reservatórios deveria levar em consideraçao, além dos aspectos hidrológicos, a minimizaçao da relaçao entre o custo operativo adicional do Setor Elétrico e o beneficio económico promovido pelo controle de possíveis cheias. Como o levantamento de tais beneficios é de difícil consecuçao, optou-se por urna decisao baseada numa alternativa de alocaçao de volume para controle de enchente que propicie a proteçao contra cheias de urna determinada garantía, e que nao acarrete apreciável aumento do risco de déficit e de geraçao térmica no Sistema Interligado.

3. Planejamento para Operaçao de Controle de Cheias

No planejamento da operaçao hidráulica dos reservatórios dos sistemas interligados brasileiros, objetivando o controle de cheias, sao consideradas duas etapas. Na primeira, chamada de prevençao, os órgaos responsáveis pela operaçao dos sistemas dotam-se de recursos físicos e materiais para o controle de enchentes.

Os recursos físicos sao os volumes vazios deixados nos reservatórios para proteçao de restriçoes de jusante, chamados de volumes de espera, para amortecimento de cheias de magnitudes até o correspondente tempo de recorrência adotado, bem como os rebaixamentos para proteçao de restriçoes a montante.

Os recursos materiais para o controle de cheias correspondem ás redes de mediçao e previsao hidrometeorológicas e sistemas de transmissao de dados, de forma que a manipulaçao dessas informaçoes forneça elementos para decisoes operativas mais adequadas dentro cada situaçao.

Em urna segunda etapa, caracterizada pela defíniçao de regras de operaçao para o controle de cheias, sao elaboradas todas as medidas que devem ser tomadas durante a ocorrência de cheias, tanto de caráter administrativo como técnico. As medidas administrativas compreendem o estabelecimento de responsabilidades nas decisoes operativas entre as diversas pessoas e órgaos envolvidos na operaçao e o acionamento de equipes para execuçao de tarefas de emergencia, tais como, o aviso ou a remoçao de moradores ribeirinhos etc.

As medidas técnicas tratam principalmente da coleta, processamento e análise dos dados operativos observados no sistema, visando a sua utilizaçao para quantifícaçao e previsao das afluências, e das decisoes operativas indicadas pelas regras de operaçao para controle de cheias estabelecidas. Estas regras sao um conjunto de instruçoes através das quais sao definidas operaçoes hidráulicas que promovam a contençao de cheias, mas que, prioritariamente, garantam a segurança do sistema hidráulico e/ou reservatório, tendo em vista as condiçoes operativas caracterizadas pelos níveis d'água nos reservatórios, afluências, e restriçoes operativas tais como: taxas de variaçao dos armazenamentos e das defluências. As regras de operaçao sao elaboradas considerando duas condiçoes: operaçao normal e operaçao em emergencia.

3.1 Prevençao de Cheias

Dentro da primeira etapa, denominada prevençao, serao aqui abordados métodos utilizados para o cálculo dos volumes de espera para controle de cheias. O primeiro método adotado é o chamado Método da Curva Volume x Duraçao, conforme Beard (1963), adaptado ao Sistema Brasileiro.

3.1.1 Método da Curva Volume x Duraçao

Este método relaciona, para cada intervalo de tempo com duraçao de d dias consecutivos, o máximo volume afluente, definido por:

onde:

va (d) - máximo volume afluente para a duraçao de d dias;
d - duraçao em dias;
q(t+j) - vazão média no día t+j;
Dt - intervalo de discretização do tempo (1 dia = 86400 s);
h - número de dias da estação chuvosa;
t - día.

Para cada local de interesse, a partir da serie histórica de vazoes naturais medias diarias e admitindo urna vazao defluente máxima que nao cause danos a jusante (descarga de restriçao), pode-se definir, para cada ano hidrológico, o volume vazio necessário para absorver cheias com qualquer duraçao. Este volume, denominado volume de espera para o ano i, pode ser representado pela seguinte expressao:

onde:

ve(i) - volume de espera para o período chuvoso do ano hidrológico i;
qr - descarga de restriçao;

A duraçao associada a este volume é chamada duraçao crítica.

A questao que se pretende resolver é a determinaçao de alternativas de volumes de espera a serem consideradas nos estudos de prevençao de cheias. Como nao se sabe qual será a seqüência de vazoes futuras, o que permitiría urna soluçao determinística para o cálculo dos volumes de espera adota-se urna soluçao probabilística.

A partir de um estudo de freqüência dos eventos máximos com varias duraçoes da s histórica de vazoes, procura-se fazer ajustes de distribuiçoes teóricas de probabilidades para c duraçao , através das quais pode-se determinar os volumes afluentes a cada duraçao associados a i probabilidade p' de ocorrência.

A Figura 1 mostra a obtençao da Curva Volume x Duraçao para a probabilidade p'.

Figura 1. - Obtengo da Curva Volume x Duraçao associada a probabilidade fixa p' de ocorrência

Neste caso, va (d) é tal que:

P [VA (d) va (d)]=p'

A questao fundamental que se coloca para a utilizaçao do método é relativa a qual distribuiçao teórica de probabilidade deve ser ajustada. Este assunto tem sido discutido em varios trabalhos conforme relatado em NERC (1975). USWRC (1977), Kite (1977) e ELETROBRAS (1987). Contudo, urna conclusao bastante difundida é que nao se dispoe de urna distribuiçao "ótima" para o ajuste a series de vazoes extremas. Portante, desejando-se urna maior segurança para o estudo, é necessário urna análise do ajustamento de diversas distribuiçoes. Em geral, a proteçao considerada no controle de cheias, para situaçoes normáis, nao vai além de urna recorrência de 50 anos. Portanto, o bom ajustamento da distribuiçao teórica á faixa amostral tem grande relevância nesta situaçao.

Uma outra dificuldade da aplicaçao do Método da Curva Volume x Duraçao é que o mesmo informa somente um único volume de espera para toda a estaçao chuvosa. Este método nao considera a variaçao do potencial de cheia com o decorrer da estaçao chuvosa. Tal soluçao, na medida em que existem interesses conflitantes no uso do aproveitamento, nao é eficiente. Realmente, em regioes com sazonalidade bem definida, o risco para o qual o volume de espera é dimensionado so se verifica no período inicial da estaçao chuvosa, decaindo conforme se aproxime o final da estaçao. Logo, a alocaçao constante do volume de espera superdimensiona a proteçao desejada. Ha que se ressaltar, aínda, que, dependendo das vazoes futuras, pode-se nao conseguir o reenchimento total do reservatório ao final da estaçao chuvosa, implicando numa reduçao das disponibilidades energéticas durante a estaçao seca que se segué.

A estrategia que vem sendo adotada para compatibilizar a manutençao do risco com a evoluçao temporal do período chuvoso, consiste na determinaçao de volumes de espera a partir da censura da serie de vazoes naturals medias diarias, conforme o avanco da estaçao chuvosa.

Em um sistema de resérvatenos, ha aínda a necessidade de se determinar a alocaçao espacial dos volumes de espera. A existência de varios locáis sujeitos á descargas de restriçao assim como a dependencia temporal de um sistema em cascata, exigem o cálculo dos volumes de espera para cada um destes locáis.

Entretanto, como nem sempre estes locáis estao imediatamente a jusante de resérvatenos capazes de alocar estes volumes, é necessário se dispor destes em resérvatenos mais a montante, conforme será visto na aplicaçao á bacía do Alto Paraná. Este procedimento pode tornar a operaçao de controle de cheias vulnerável á distribuiçao de cheia na bacia.

Visando superar as dificuldades apontadas para a alocaçao temporal e espacial de volumes de espera, o setor elétrico brasileiro, através do Centro de Pesquisas de Energía Elétrica, CEPEL, vem realizando estudos de novas alternativas metodológicas de cálculo e alocaçao de volume de espera. Ja existe hoje desenvolvido o Método das Trajetórias Críticas que, associado á Teoría das Condiçoes de Controlabilidade, permite a determinaçao, para um sistema de reservatórios, dos volumes de espera necessários e suficientes para o controle de cheias em todo o sistema.

3.1.2 Metodología Alternativa

· Método das Trajetórias Críticas

Este método foi desenvolvido no CEPEL por Kelman (1987) e permite determinar para cada dia da estaçao chuvosa, o volume de espera associado a urna probabilidade de ocorrência p', previamente estabelecida.

O método emprega um algoritmo recursivo sobre as vazoes naturals medias diarias do local a ser estudado. Partindo-se de um volume de espera nulo no final do último dia do período chuvoso, no dia h de um ano i, da vazao media diaria neste dia e da descarga de restriçao, determina-se o volume no inicio deste dia da seguinte forma:

(4)

onde:

ve - volume de espera;
h - índice do último dia do período chuvoso;
i - Índice do ano;
q(h,i) - vazao media diaria no dia h do ano i;
qr - descarga de restriçao
_t - intervalo de discretizaçao (1 dia = 86400 s).

Ao inicio do penúltimo dia, o volume de espera deverá ser igual a:

(5)

De urna forma geral, entao, dispondo-se sempre do volume necessário no inicio do dia posterior, pode-se, recursivamente, determinar-se os volumes necessaries nos dias anteriores:

(6)

A evoluçao do volume de espera ao longo do tempo para o i-ésimo ano denomina-se trajetória crítica. A figura 2 a seguir, aprésenla a trajetória crítica para o ano i, mostrando as regioes segura e insegura para a operaçao de controle de cheias do reservatório.

Figura 2. - Evoluçao do Volume de Espera para o i-ésimo ano.

Suponha, agora, 3 anos distintos, cujas trajetórias estao na figura 3. A envoltória das mesmas, representada em linha tracejada, garantirá que em nenhum instante, nao importando o ano ocorido, a vazao defluente será superior a descarga de restriçao.

Figura 3. - Definiçao da envoltória das trajetórias críticas.

Em urna serie de n anos, conseqüentemente n trajetórias, a envoltória garantirá a nao ocorrência de violaçao da descarga de restriçao.

A envoltória é definida por:

Como a seqüência de vazoes futuras nao é conhecida, adota-se, assim como no caso da Curva Volume x Duraçao, urna abordagem probabilística, sendo que neste caso o universo de possíveis seqüência s de vazoes futuras é obtido através de series sintéticas de vazoes. A utilizaçao de series sintéticas é imperiosa pois as series históricas de vazoes diarias disponíveis nos locais de interesse para os estudos de prevençao de cheias geralmente nao ultrapassam 60 anos, período este, ainda pequeño para a determinaçao de trajetórias associadas aos tempos de retomo comumente adotados pelo setor elétrico brasileiro.

Como, provavelmente, a alocaçao da envoltória obtida das séries sintéticas de vazoes seria inviável, é necessário a aceitaçao de alguma probabilidade p' de insucesso no controle de cheias. Defíne-se, entao, um conjunto de k trajetórias críticas que nao serao consideradas na determinaçao da envoltória de volumes de espera.

Durante os estudos e testes realizados até entao com o método das Trajetórias Críticas, o modelo de geraçao de series sintéticas utilizado tem sido o modelo DIANA, desenvolvido por Kelman (1986).

· Teoría das Condiçoes de Controlabilidade

O Método das Trajetórias Críticas resolve a questao da alocaçao temporal dos volumes de espera, entretanto, para um sistema de resérvatenos é necessário, ainda, a determinaçao de como este volume deve estar disponível espacialmente.

Os problemas de sistemas multireservatórios para controle de cheias foram analisados teóricamente pela primeira vez por Manen (1984), levando a formulaçao da teoría das condiçoes de controlabilidade para a análise determinística de sistemas de reservatórios com apenas um local sujeito á restriçao. Damázio (1989) estendeu estes resultados de forma a considerar também sistemas com múltiplos locáis sujeitos á restriçao, agregando ainda á esta teoria a naçao de envoltórias, permitindo um tratamento estocástico do problema.

Tomando como exemplo um sistema de dois reservatórios, R₂ e R₁ em série, sendo R₂ o reservatório de montante, no qual existem restriçoes de vazao máxima imediatamente a justante de cada reservatório. O cálculo do volume de espera a ser alocado em cada reservatório e em cada dia deve considerar que existe um mínimo espaço vazio a ser alocado em R₂, funçao exclusiva de sua restriçao e das afluências a R₂. Por outro lado, o mínimo espaço vazio a ser alocado á jusante depende do total afluente á R₁ (soma da afluência á R₁ com a afluência incremental entre R₂ e R₁) e também do espaço vazio que tiver sido alocado em R₂.

Para um dia qualquer t, o dominio das soluçoes possíveis é caracterizado pelas seguintes equaçoes, chamadas de condiçoes de controlabilidade:

(1)
(2)
(3)

onde: E(t) é o espaço vazio e ve(t) é o limite inferior que se deseja calcular.

Os valores de ve(t) servem para definir a regiao viável para os espaços vazios no dia t. Em geral, para qualquer sistema multireservatório a regiao viável será sempre caracterizada por limites inferiores ve(t) para somas de espaços vazios de reservatórios pertencentes aos varios subsistemas existentes neste sistema.. Os subsistemas do sistema de dois reservatórios descritos ácima sao {R₁, {R₂} e {R₁, R₂}. A figura 4 ilustra a fronteira entre as regioes viável e inviável para o controle de cheias neste sistema como um todo para um determinado dia.

Damázio (1989), mostra como os limites inferiores de espaços vazios, ve(t) podem ser obtidos para cada subsistema.

Figura 4. - Regioes viáveis e inviáveis para os espaços vazios num determinado dia

3.2 Regras de Operaçao para Controle de Cheias

Urna vez escolhida a alternativa de alocaçao de volume de espera, passa-se á segunda etapa do planejamento da operaçao hidráulica onde podem ser visualizadas duas condiçoes, a saber: operaçao normal e operaçao em emergencia.

· Operaçao Normal

Esta operaçao pode ser caracterizada pela ocorrência de urna cheia que permanece, ao longo de sua duraçao , sob o controle do Centro de operaçao, nao havendo perspectivas de esgotamento dos volumes de espera dos resérvatenos nem da liberaçao de descargas defluentes que ultrapassem restriçoes ajusante dos aproveitamentos.

Durante a operaçao normal para controle de cheias, sao recebidas e processadas as informaçoes de dados provenientes de postos hidrometeorológicos da rede de mediçao, oferecendo um quadro aproximado da situaçao real do sistema e permitindo urna boa confiabilidade das decisoes.

É indispensável ainda a disponibilidade de regras de operaçao que recomendem os valores das descargas totais em funçao de varios parámetros, tais como níveis em pontos críticos dos rios, vazoes afluentes em pontos situados a montante ou a jusante dos reservatórios, vazoes dos principais afluentes, alturas de chuva etc. Estas regras sao frutos de urna longa experiencia operativa ou podem ser obtidas de simulaçoes.

As regras básicas adotadas em urna operaçao normal sao:

1. enquanto a vazao afluente for inferior ao limite da restriçao de defluência, a descarga defluente total (vertida + turbinada) deve ser igual á vazao afluente. Em conseqüência , o nivel do reservatório deve se manter no valor recomendado do volume de espera;

2. quando a vazao afluente for superior ao limite da restriçao de defluência, a descarga defluente total deve ser igual ao limite da restriçao de defluência. Em conseqüência , o nivel do reservatório começa a subir, iniciando-se a utilizaçao do volume de espera e o amortecimento da cheia;

3. tendo sido o volume de espera ocupado parcial ou totalmente e iniciando-se o seu esvaziamento, a descarga defluente total deve ser mantida igual ao valor limite da restriçao de defluência, de forma a restabelecer o mais prontamente possível o volume de espera, sem provocar danos ajusante;

4. próximamente ao restabelecimento do volume de espera, a descarga defluente total deve ser reduzida do limite da restriçao de defluência para o valor da afluência, através de urna taxa de reduçao de defluência que nao provoque danos ajusante.

· Operaçao em Emergencia

Esta operaçao pode ser caracterizada pela ocorrência de urna cheia com perspectivas de esgotamento dos volumes de espera dos reservatórios, sendo necessário providenciar descargas defluentes totais que superam as limitaçoes impostas por eventuais restriçoes, ocasionando situaçoes críticas e provocando danos.

Outra situaçao possível que caracteriza a condiçao de emergencia é a perda de comunicaçao da usina com o Centro de operaçao. Neste caso, o elemento encarregado da operaçao da usina deve estar autorizado para tomar as providencias cabíveis, para as quais deve estar disponível urna sistemática de procedimentos bem definidos, que garantam a segurança da barragem.

Em ambas as situaçoes, é indispensável a existencia de regras de operaçao que indiquem, a cada instante, qual defluência deve ser programada de forma a garantir a segurança da estrutura das barragens, sem provocar enchentes mais críticas do que as que ocorreriam sob condiçoes naturais de escoamento. Um exemplo desta regra é o diagrama de operaçao em emergencia apresentado no Anexo 1.

· Diagrama de Operaçao em Emergencia

Urna ferramenta bastante utilizada, durante a ocorrência de urna cheia, para obtençao do valor da defluência mínima necessária em cada aproveitamento, é o denominado diagrama de operaçao em emergencia apresentado por U.S. Army - Corps of Engineers (1959), Beard (1963), CECCA (1977). Este diagrama define a descarga que deve ser defluída, quando sao conhecidos os níveis do reservatório e a taxa de variaçao dos níveis ou a afluência, tendo em vista urna previsao do mínimo volume afluente que pode ser esperado em urna particular cheia. A hipótese admitida é de que o hidrograma tenha atingido o seu pico para o valor atual e de que os valores futuros ocorrerao segundo urna lei exponencial que define a recessao das vazoes de um rio. O volume mínimo afluente a ser esperado pode ser calculado pela área sob o trecho de descida do hidrograma. A partir deste volume afluente mínimo esperado, com urna dada afluência presente e o nivel atual do reservatório, pode ser calculada a descarga defluente, de modo que seja utilizado o volume do reservatório até a sua capacidade disponível, limitada no seu nivel máximo normal ou, quando possível, pela sua curva de sobrecarga induzida. Uma familia de curvas características de afluências, como a mostrada no Anexo 1, pode ser traçada a partir destes cálculos.

· Curva de Sobrecarga Induzida

No esquema a ser desenvolvido para operaçao em emergencia, salienta-se o conceito de curva de sobrecarga induzida. A sobrecarga induzida é obtida quando as comportas do vertedor sao operadas todas ao mesmo tempo, através de aberturas parciais, de modo a possibilitar o armazenamento do excesso de afluência em relaçao á descarga defluente total. A sobrecarga induzida, para cada abertura parcial das comportas, está limitada pela cota do topo das comportas do vertedor para esta mesma abertura, de modo a evitar o galgamento das mesmas.

4. Aplicaçao á Bacía do Alto Paraná

A bacia do Alto Parana drena parte das regioes Sudeste, Centro-Oeste e Sul do Brasil, conforme mostra o Anexo 2. Está localizada no eixo de maior desenvolvimento do país, envolvendo além do Distrito Federal, áreas dos estados de Goiás, Mato Grosso do Sul, Minas Gerais, Sao Paulo, Paraná e Santa Catarina.

Tendo em vista as diversidades hidrológicas, bem como a topología dos aproveitamentos, pode-se considerar, do ponto de vista do controle de cheias, tres sistemas distintos:

1. Bacia do Alto Paraná até Jupiá;
2. Bacia do rio Paranapanema;
3. Bacia do rio Iguaçu;

A bacia do Alto Paraná até Jupiá, com 470.000 km² de área de drenagem, abrange além do trecho do rio Parana, os rios Paranaíba, Araguari, Grande, Pardo e Tietê. Nestes rios estao implantados 27 aproveitamentos hidroelétricos pertencentes as empresas FUMAS, CESP, CEMIG e CELG, sendo que 14 possuem reservatórios, nos quais sao alocados volumes de espera.

A bacia do rio Paranapanema possui urna área de drenagem de cerca de 100.000 km². Todos os 6 aproveitamentos hidroelétricos existentes pertencem á CESP, sendo que 3 deles possuem reservatórios, utilizados no controle de cheias.

A bacia do rio Iguaçu, com urna área de drenagem de 68.000 km², possui, até o momento, 5 aproveitamentos, sendo 4 dispostos em cascata no rio Iguaçu, e um de menor dimensao no rio Chopim. Estes aproveitamentos pertencem as empresas COPEL e ELETROSUL, sendo que 2 deles possuem reservatórios os quais sao utilizados também para controle de cheias.

No rio Paraná, a montante da confluencia com o rio Iguaçu, localiza-se o aproveitamento binacional de Itaipu, com urna área de drenagem da ordem de 820.000 km².

O Anexo 3 aprésenla um diagrama esquemático dos aproveitamentos existentes na bacia do rio Paraná.

O conjunto de aproveitamentos a montante de Jupiá se constituí no principal sistema, tendo em vista a sua dimensao e complexidade da sua operaçao.

4.1 Restriçoes Hidráulicas Existentes na Bacia

A bacia do rio Paraná possui diferentes tipos de restriçoes hidráulicas a serem protegidas com medidas preventivas de controle de cheias, descritas de forma resumida a seguir.

Iniciando pelo rio Grande, tem-se urna restriçao de defluência de 400 m³/s no conjunto Camargos/Itutinga, condicionada as enchentes na cidade de Ribeirao Vermelho, 190 km ajusante. Em particular, ocorrem inundaçoes nesta cidade com vazoes superiores a 1.000 m³/s, compostas pela soma das defluências das usinas ás contribuiçoes da bacia incremental, representadas principalmente pelas vazoes dos rios Capivari e das Mortes.

A jusante de Fumas tem-se urna vazao de restriçao de 4.000 m³/s devido a um porto de balsas em Sao Joao Batista do Gloria, ranchos ribeirinhos e portos de areia.

A usina de Mascarenhas de Moraes, pelo posicionamento de seus vertedouros, nao pode defluir vazoes superiores a 4.400 m³/s sem correr o risco de inundaçao da sua casa de força. Operando de forma conjugada com a usina de Estreito, ja que seu canal de fuga sofre influencia do remanso desse reservatório, a sua restriçao pode ser aumentada até 6.000 m³/s. Possui, também, restriçao de nivel mínimo no seu reservatório (665,00 m = 89,2% V.U.) em funçao das tomadas para captaçao de agua de usinas açucareiras na área do lago.

As usinas de Estreito e Jaguara possuem urna restriçao de 4500 m³/s em funçao de inundaçao em trecho da rodovia BR-050, entre as cidades de Uberaba (MG) e Ribeirao Preto (SP). Por outro lado, Volta Grande deve respeitar 5.000 m³/s para evitar a inundaçao de casas de campo ao longo do rio, enquanto Porto Colombia tem como limite 7.000 m³/s para nao causar problemas para a estrutura da ponte Gumercindo Penteado, da BR-364.

O último aproveitamento do rio Grande com restriçoes operativas é Marimbondo, de FUMAS. Possui urna restriçao de nivel máximo (445,73 m = 95% V.U.) que visa manter o nivel na ponte Gumercindo Penteado, da BR-364, na cota 446,40 m, ou seja, 50 cm abaixo do seu tabuleiro, evitando o acumulo de plantas aquáticas, o represamento de ilhas flutuantes, obstruçoes ao fluxo de agua entre seus pilares e á montante da viga lateral, e as sucessivas submersoes e emersoes da viga da ponte. Para jusante, Marimbondo possui urna descarga de restriçao de 8.000 m³/s, que atinge áreas urbanizadas da própria usina, ranchos de pescadores e causa a erosao do acesso a ponte rodoviária da BR-153.

No rio Paranaíba, o aproveitamento de Emborcaçao possui urna vazao de restriçao de 5.000 m³/s, a partir da qual existe a inundaçao da estrada de acesso á própria usina.

A usina de Nova Ponte, no rio Araguarí, afluente do rio Paranaíba, tem sua restriçao variável, da ordem de 2.000 m³/s, em funçao da afluência máxima de 2.500 m³/s de aduçao dos dois túneis de desvio da obra da usina de Miranda, 80 km ajusante de Nova Ponte.

A usina de Itumbiara passou, a partir da cheia de 1983, a operar para controlar a vazao de 7.000 m³/s, visando evitar a inundaçao de áreas de benfeitorias da cidade de Itumbiara, localizada 13 km ajusante da barragem.

Finalmente, o aproveitamento de Sao Simao, nesse mesmo rio, possui urna restriçao de 16.000 m³/s referente a ponte da rodovia BR-365 na fronteira dos estados de Minas Gerais e Goiás. Vale ressaltar que tal restriçao corresponde a urna vazao de recorrência bastante elevada.

No rio Tietê, tem-se o aproveitamento de Barra Bonita com urna restriçao, bastante severa, de 1.800 m³/s, a qual interrompe a navegaçao fluvial por impedimento do funcionamento da eclusa.

No rio Paraná, a restriçao de Jupiá, cujo valor é 16.000 m³/s, é a mais severa de toda a bacia, haja visto o grande impacto causado pelo seu rompimento, provocando inundaçoes nas cidades de Tres Lagoas, Panorama e Presidente Epitácio, e o seu baixo período de retorno, de aproximadamente 2 anos. Portante, em condiçoes naturais, na ausencia do sistema de reservatórios de montante, o risco de ocorrerem vazoes iguais ou superiores á restriçao seria de 50% a cada ano.

Na bacia do rio Paranapanema, todos os aproveitamentos pertencem á empresa CESP. Esta bacia possui um regime de vazoes misto, isto é, ora influenciado pela sazonalidade da regiao Sudeste, ora pelo regime dos rios da regiao Sul.

Em Jurumirim, defluências da ordem de 800 m³/s acarretam o inicio do galgamento e conseqüente inundaçao da antiga usina de Piraju, imediatamente ajusante. Vazoes de 1.200 m³/s trazem problemas na própria bacia de dissipaçao de Jurumirim.

Na usina de Chavantes, a vazao de restriçao adotada é de 1.800 m³/s, haja visto que valores superiores interrompem o tráfego na ponte ferroviaria existente ajusante. Com 2.200 m³/s surgem problemas na bacia de dissipaçao da usina e inicia-se o alagamento do seu núcleo administrativo. As soluçoes de relocaçao das pontes e instalaçoes da CESP sao de custo elevado, exigindo a alocaçao de volumes de espera e outras medidas para controlar a restriçao de 1.800 m³/s, cujo período de retorno é de 6 anos.

Finalmente, em funçao das últimas cheias, é decretada "Situaçao de Emergencia" nos municipios de Marilena, Sao Pedro do Paraná, Porto Rico, Querencia do Norte e Umuarama sempre que as vazoes em Porto Sao José, no rio Paraná, junto a foz do Paranapanema, excederem 21.000 m³/s.

Neste caso, para controlar as vazoes, a usina de Capivara passa a ser operada, temporariamente, com urna restriçao condicionada, de modo que nao seja excedido os 21.000 m³/s, resultante da soma das vazoes dos rios Paraná e Paranapanema.

No rio Iguaçu o aproveitamento de Foz do Areia, o primeiro da cascata, possui urna restriçao de operaçao a montante, caracterizada pela influencia, para certas condiçoes, nos níveis de inundaçao da cidade de Uniao da Vitória, localizada na cabeceira do seu reservatório. A influencia do nivel deste reservatório sobre o nivel do rio Iguaçu, em Uniao da Vitória, é explicada pelas curvas de remanso determinadas pela COPEL. Assim, para cada situaçao de afluência, o nivel de restriçao a ser considerado em Foz do Areia pode ser determinado em tempo real, com base nos níveis da agua observados, simultáneamente, em Foz do Areia, Porto Vitória e Uniao da Vitória.

Salto Santiago e Salto Osório, ajusante de Foz do Areia, operam condicionando suas defluências as contribuiçoes incrementáis observadas até a confluencia com o rio Chopim, de forma a respeitar o valor de 17.000 m³/s nessa seçao para evitar danos á casa de força da usina Julio de Mosquita Filho (JMF). A usina de Salto Santiago possui, ainda, restriçao de liberaçao de vazoes superiores a 19.000 m³/s, para evitar a inundaçao de sua própria casa de força.

4.2 Séries de Vazoes Disponíveis na Bacía

As séries de vazoes naturais medias diarias nos locáis de interesse para os estudos de prevençao de cheias tem inicio de observaçao variável. As series dos postos a montante de Jupiá foram estendidas até 1994, utilizando-se métodos de reconstituiçao de vazao natural baseados em balanço hídrico de resérvatenos e/ou correlaçao entre postos fluviométricos.

4.3 Aplicaçao da Metodología

A metodología adotada para determinaçao dos volumes de espera para a bacia do rio Paraná é da Curva Volume x Duraçao, descrita no item 3.1 deste relatório. Esta metodología vem sendo aplicada nesta bacia desde 1977. A única exceçao foi a utilizaçao do método das Tajetórias Críticas para o cálculo dos volumes de espera próprio em Nova Ponte, para a proteçao da obra UHE Miranda.

A variaçao temporal dos volumes de espera a serem alocados nos resérvatenos desta bacia, a montante de Jupiá, foi obtida pela censura continua das series de vazoes naturais diarias, considerando, para o cálculo dos volumes de espera, series cada vez menores, conforme o avanço da estaçao chuvosa.

Na análise de freqüêcias, a distribuiçao de probabilidades adotada na determinaçao dos volumes em todos os locáis foi a Log-Pearson III. Esta metodologia determina o volume vazio necessário para a proteçao de cada local sujeito a urna restriçao de vazao, denominados volume de espera próprio. Desta forma, foi aplicada a metodologia para os postos de Camargos, Fumas, Mascarenhas de Moraes, Jaguara, Volta Grande, Porto Colômbia, Marimbondo, Emborcaçao, Itumbiara, Sao Simao, Barra Bonita e Jupiá.

Os estudos de volume de espera para prevençao de cheias elaborados pelo GCOI normalmente contemplam varias alternativas de riscos. A faixa de valores de tempo de retorno usualmente considerada varia de 25 a 50 anos, exceto no caso de determinados locáis, onde o impacto do rompimento da restriçao é tao grande, como a perda da própria usina, que prefere-se elevar o grau de proteçao aumentando o valor do tempo de retorno a ser adotado.

Na bacia do rio Paranapanema os reservatórios da CESP tiveram seus volumes de espera dimensionados para cheias correspondentes a recorrência de 100 anos. Estes valores sao alocados todo ano em funçao do regime de cheias conjunto das regioes Sudeste e Sul do Brasil.

A bacia do rio Iguaçu aprésenla, atualmente, duas restriçoes que devem ser consideradas para o cálculo de volumes de espera, quais sejam: as casas de força de Salto Santiago e de Julio de Mesquita Filho. Os volumes de espera definidos para proteger as restriçoes ácima sao alocados de forma independente, pois os tempos de recorrência associados aos volumes nao sao iguais. Portante, em Salto Santiago fica disponível um volume de espera para proteger a sua restriçao própria mais o volume necessário para proteger Julio de Mesquita Filho.

4.4 Resultados dos Volumes de Espera

Na bacia do Alto Paraná o caso mais crítico corresponde ao aproveitamento de Jupiá, pois está sujeito á restriçao mais severa da bacia e, por ser a fío d'água, nao pode alocar volume de espera para controlar suas defluências.

Tem-se, entao, que repartir o volume de espera necessário para proteçao da restriçao deste aproveitamento entre os demais resérvatenos da bacia, além de se planejar urna operaçao conjugada de controle de cheias. Com exceçao dos aproveitamentos a fio d'água e dos reservatórios de Camargos e Barra Bonita, cujas operaçoes de controle de cheias visam exclusivamente a proteçao das restriçoes imediatamente a jusante dos mesmos, todos os demais aproveitamentos situados a montante de Jupiá participam desta repartiçao.

Considera-se que urna boa repartiçao do volume de espera calculado para Jupiá, do ponto de vista do controle de cheias, será aquela que mantiver a maior parcela desses volumes o mais próximo possível de Jupiá. O criterio que vem sendo adotado desde 1982, baseado em urna relaçao linear com as medias das vazoes máximas mensais do histórico, nos meses da estaçao chuvosa (dezembro a abril), fornece urna distribuicao segura para a operaçao integrada de controle de cheias nesta bacia, pois quanto mais a jusante o aproveitamento se localizar, maior será sua media e, conseqüentemente, maior será a parcela que lhe caberá do volume de Jupiá.

Para efeito de reparticao dos volumes de espera destinados a proteçao de Jupiá, considera-se que os reservatórios de Fumas e Mascarenhas de Moraes formam um único reservatório, cuja vazao dos meses da estacao chuvosa é igual a vazao media em Mascarenhas de Moraes.

Para qualquer tempo de recorrência estudado, os volumes de espera de Jupiá, quando distribuidos pelos aproveitamentos de montante, sempre superam os volumes de espera necesários para o próprio aproveitamento, os denominados volumes próprios, que sao calculados para atender específicamente as restriçoes existentes a jusante de cada aproveitamento.

No entanto, no caso das restriçoes a jusante de Mascarenhas de Moraes, Itumbiara e Nova Ponte, que sao respectivamente a casa de força da usina, a cidade de Itumbiara e a obra da usina de Miranda, decidiu-se, para urna maior proteçao, considerar a alocaçao nestes reservatórios da soma das parcelas decorrentes da distribuicao do volume de espera de Jupiá com os seus respectivos volumes de espera próprios.

Os aproveitamentos de Jaguara, Volta Grande e Porto Colombia, por serem a fio d'água, têm a proteçao de suas restriçoes exercida pela operaçao dos reservatórios de Fumas e Mascarenhas de Moraes, os quais possuem restriçoes próprias compatíveis com as dos referidos aproveitamentos.

Foram levantadas alternativas de alocaçao de volumes de espera, no sistema de reservatórios a montante de Jupiá, correspondentes a varios tempos de recorrência. Foi feita urna avaliaçao dos impactos energéticos no Sistema Interligado Sul/Sudeste, decorrente da adoçao de cada alternativa, com a finalidade de se obter indicadores significativos, tais como: riscos de déficit, custos de geraçao térmica e de déficits e probabilidades de enchimento dos reservatórios no final da estacao chuvosa. A alternativa aprovada para a estacao chuvosa 1995/1996 para proteçao da restriçao de Jupiá corresponde a um tempo de recorrência de 35 anos, nao se alterando em relacao aos últimos anos.

O volume útil total desses reservatórios a montante de Itaipu é da ordem de 98 bilhoes de metro cúbicos, dos quais sao reservados cerca de 20% para volume de espera, durante a estaçao chuvosa, correspondente aos meses de dezembro a abril. Este valor se reduz nos meses de maio a novembro pois, neste período, so estao alocados volumes de espera nos reservatórios da bacia do rio Paranapanema.

4.5 Regras de operaçao

No caso da bacia do rio Paraná, deve-se destacar a existencia de dois tipos de reservatórios quanto á natureza de suas operaçoes de controle de cheias:

· reservatórios para a operaçao integrada na bacia;
· reservatórios para a operaçao isolada na bacia.

Os reservatórios de Camargos e de Barra Bonita, devido as suas localizaçoes e aos seus recursos físicos, sao considerados reservatórios para a operaçao isolada de controle de cheias, cujo único objetivo direto é a proteçao de suas restriçoes próprias, da mesma forma que os reservatórios da bacia do rio Paranapanema.

Os demais reservatórios a montante de Jupiá operam visando a proteçao nao so de suas restriçoes próprias como também da restriçao de Jupiá. Estes reservatórios compoem o sistema de reservatórios para a operaçao integrada de controle de cheias da bacia do rio Paraná.

Para efeito de urna operaçao integrada de controle de cheias para a proteçao da restriçao de Jupiá sao considerados os seguintes subsistemas: Paranaíba; Grande; Ilha Solteira/Três Irmaos e Tietê. Estes subsistemas foram divididos em segmentos. Quando a afluência natural ao sistema até Jupiá ultrapassa a sua descarga de restriçao, os volumes de espera dos subsistemas passam a ser utilizados, de maneira conjugada, mantendo-se a defluência de Jupiá na sua restriçao. A seqüência de utilizacao dos subsistemas procura se aproximar de um enchimento em paralelo, por segmentos, dos volumes de espera do sistema.

Os volumes de espera a serem alocados tiveram sua evoluçao discretizada em patamares semanais.

Quando os reservatórios encontram-se com seus armazenamentos inferiores ao nivel meta, devem ser utilizados os "Diagramas de operaçao Normal", que em funçao da previsao de vazao afluente e do volume vazio disponível indicam as defluências necessárias para atingir os níveis de volume de espera. Desta forma, pode-se antecipar o inicio da evoluçao da vazao defluente até o valor de sua restriçao, evitando-se com isto as rápidas variaçoes das defluências, que nem sempre sao possíveis devido á limitaçao das taxas de máxima variaçao diaria.

Se a afluência natural ao sistema até Jupiá for maior que a sua descarga de restriçao, mesmo existindo volume de espera no sistema, deverá ser sempre consultado o Diagrama de operaçao em Emergencia do Sistema Equivalente em Jupiá, para saber se ainda é possível manter-se a defluência deste aproveitamento na sua restriçao ou se a mesma deverá ser rompida. Para se obter esta defluência do diagrama de Jupiá deve-se dispor dos dados da afluência natural verificada e do volume de espera disponível no sistema para Jupiá.

Na fase da operaçao em tempo real funciona urna Coordenaçao Operacional para Controle de Cheias, a qual é exercida pelo CNOS - Centro Nacional de operaçao de Sistemas da ELETROBRAS, cujas principais atribuiçoes sao:

· coordenar a operaçao integrada dos reservatórios durante o período especificado, aplicando as regras de operaçao aprovadas;

· coordenar o programa de defluências e a seqüência de enchimento e deplecionamento dos reservatórios em funçao do estado hidráulico do sistema;

· garantir facilidades e a confiabilidade na troca de informaçoes de dados hidráulicos e hidrológicos entre as empresas envolvidas.

5. Análise da Experiencia Operativa

A partir de 1977 o setor elétrico brasileiro começou a incorporar, de forma sistematizada, os estudos de controle de cheias no planejamento e programaçao da operaçao dos reservatórios dos Sistemas Interligados Sul/Sudeste e Norte/Nordeste. Inúmeros eventos de cheias significativas ocorreram ao longo destes anos, em varias bacias que permitiram avaliar o desempenho da estrategia de controle de cheias utilizada pelo GCOI. Nesta direçao, selecionamos o ano de 1992, que apresentou ao longo do seu transcurso, varios eventos significativos de cheias em diversas regioes brasileiras, particularmente para a bacia do alto rio Paraná. Por oportuno, devido a indivisibilidade da bacia, incluímos, também, urna análise resumida da operaçao dos reservatórios do rio Iguaçu, bem como suas implicaçoes com a operaçao hidráulica realizada ajusante de Itaipu nos meses de maio e junho de 1992.

5.1 Operaçao de Controle de Cheias na Bacia do Alto Paraná: Jan/Fev -1992

Chuvas intensas e de longa duraçao verificadas na regiao Sudeste a partir da 2^a quinzena de Janeiro de 1992, decorrentes de urna frente estacionaria durante quase dez dias, provocaram cheias generalizadas na bacia do Alto Paraná, mais notadamente nas cabeceiras do rio Grande. A precipitaçao total neste mês de Janeiro chegou a alcançar até 200% da media histórica em alguns postos da cabeceira do rio Grande.

Na usina de Camargos, no alto rio Grande, em 23.01.92 ocorreu a maior cheia de um histórico de 60 anos registrado neste local, no valor de 1.630 m³/s. A liberaçao de urna defluência máxima de 1.170 m³/s em Camargos evitou a destruiçao de urna ponte na estrada BR-265, embora tenha causado urna interrupçao temporaria no transito.

Este abatimento da cheia efetuado por Camargos nao foi suficiente para evitar danos á localidade de Ribeirao Vermelho, pois so as vazoes incrementáis, nao controladas pela operaçao do reservatório, chegaram a atingir 4.500 m³/s, muito ácima da restriçao da cidade, 1.100 m³/s.

Para os demais reservatórios da cascata a operaçao de controle de cheias realizada foi bastante eficiente.

No reservatório de Fumas, em 3.01.95, foi registrada a vazao media diaria mais elevada do histórico de 62 anos: 9.000 m³/s. Apesar deste quadro desfavorável, as vazoes defluentes foram totalmente controladas, atingindo-se a restriçao de defluência (4.000 m3/s) apenas a partir de 26 de Janeiro, conforme mostrado no Anexo 4.

O sucesso da operaçao de controle de cheias dos reservatórios da bacía do rio Grande (Fumas, M. Moraes, Marimbondo e Agua Vermelha), permitiu que fossem evitados os seguintes prejuízos:

· inundaçao da usina acucareira de Passos;
· interrupçao do tráfego de balsa em Sao Joao Batista do Gloria;
· paralisaçao do tráfego na ponte da BR-050 no trecho Uberaba-Ribeirao Preto;
· paralisaçao das atividades extrativas em portos de areia;
· inundaçao da casa de força da usina Mascarenhas de Moraes;
· interrupçao do tráfego da BR-364 na ponte Gumercindo Penteado.

Mais ainda, o controle de cheias no rio Grande evitou que fossem desalojados diversas familias em áreas ribeirinhas, a perda de culturas agrícolas e a dizimaçao de rebanhos.

Na bacía do rio Paranaíba, o evento mais significativo da estacao chuvosa ocorreu no dia 11.02.92., conforme pode ser observado também no Anexo 4. A operaçao de controle de cheias do reservatório de Itumbiara permitiu manter as vazoes defluentes em no máximo 6.750 m³/s, respeitando a restriçao para proteçao da cidade de Itumbiara, que é de 7.000 m³/s.

Apesar das chuvas mais intensas nas bacias dos rios Grande e Paranaíba nao terem ocorrido simultáneamente, no período de 20 de Janeiro a 15 de fevereiro, as vazoes naturais levantadas para Jupiá neste período foram expressivas: Em torno do dia 24 de Janeiro, logo após a cheia ocorrida no rio Grande, a vazao natural em Jupiá atingiu o valor de 25.400 m³/s, entrando em recessao até 2 de fevereiro. Logo após esta data, em funçao do aumento das vazoes no Paranaíba, as vazoes naturais em Jupiá voltaram a subir, alcanzando cerca de 24.300 m³/s, conforme apresentado no Anexo 5. Esta cheia em Jupiá está entre as seis maiores de um histórico de 70 anos, em termos de máxima vazao media diaria.

Portante, caso nao houvesse urna eficiente operaçao de controle de cheias nos subsistemas Grande e Paranaiba, bem como do reservatório de Ilha Solteira, as localidades de Tres Lagoas, Panorama e Presidente Epitácio, a jusante da usina de Jupiá, sofreriam inundaçoes. A operaçao coordenada realizada permitiu que em ambos os eventos a defluência de Jupiá nao ultrapassasse a restriçao de 16.000 m3/s, significando urna acumulaçao no sistema de reservatórios brasileiros da ordem de 13 bilhoes de metros cúbicos, ou seja aproximadamente 13% de sua capacidade total.

Deve-se destacar no primeiro evento que, para se elevar a defluência em Jupiá, foi necessário que a CESP e a Defesa Civil removessem cerca de 100 familias residentes nos locáis atingidos. Por isso, a vazao defluente foi mantida em 6.500 m³/s até o dia 19 de Janeiro, sendo gradativamente aumentada até 15.000 m3/s no dia 27. A partir daí, em funçao da melhoria das condiçoes meteorológicas, a coordenaçao do controle de cheias manteve esse valor de defluência, permitindo o esvaziamento dos volumes de espera ocupados na bacia do rio Grande.

5.2 Operaçao de Controle de Cheias na bacia do Rio Iguaçu: Mai/Jun - 1992

A bacia do rio Iguaçu localiza-se na regiao Sul do país, em grande parte do estado Paraná, tendo também áreas de contribuiçao em Santa Catarina e na República Argentina.

Essa bacia possui certas características que determinam distintos tipos de enchentes ao longo da mesma. O trecho conhecido como Alto Iguaçu, a montante da localidade de Porto Vitória, caracteriza-se por declividades pequenas. Neste trecho, os escoamentos evoluem de forma gradual e permitem boas previsoes de vazoes. Por outro lado, no trecho ajusante de Porto Vitória, ocorre um aumento acentuado das declividades, provocando respostas rápidas as precipitaçoes, com reduçao do horizonte de previsao.

As fortes chuvas ocorridas na regiao Sul em maio/junho de 1992 elevaram as vazoes no rio Iguaçu e causaram a segunda maior cheia dos últimos 60 anos. O total de precipitacao acumulado em maio de 1992 chegou a atingir 500 mm para algumas localidades, valor este cinco vezes superior ao total medio histórico.

Na regiao a montante de Foz do Areia, a vazao máxima registrada na cidade de Uniao da Vitória foi de 3.800 m³/s, causando inundaçoes para as populaçoes ribeirinhas. Vale ressaltar que quando da ocorrência deste evento máximo nao havia influencia do remanso do reservatório de Foz do Areia naquela cidade, devido ao deplecionamento efetuado.

Nesta época, o aproveitamento de Segredo, da COPEL, situado imediatamente a jusante de Foz do Areia, estava em construçao, e propiciou um grande abatimento da cheia, devido a limitaçao de descarga imposta pelos seus dois tunéis de desvio. Como conseqüência da grande acumulaçao ocorrida, a ensecadeira da obra de Segredo foi galgada, sendo necessário se implementar diversas medidas emergenciais.

Devido a magnitude da cheia, o volume de espera alocado em Salto Santiago, da ELETROSUL, nao foi suficiente para garantir a vazao de restriçao de 17.000 m3/s na seçao correspondente a usina Julio de Mesquita Filho, da COPEL, ocorrendo a inundaçao desta usina. No entanto, a defluência de Salto Santiago nao ultrapassou a sua descarga de restriçao (19.000 m3/s), evitando a inundaçao dessa usina, fundamental para o atendimento energético da regiao Sul.

Durante o desenrolar da cheia na bacia do rio Iguaçu, por volta das 19 horas do dia 29.05, a ELETROBRAS, que realizava o seu acompanhamento e controle, através do CNOS, em conjunto com a COPEL e ELETROSUL, foi informada pelo Ministerio das Relaçoes Exteriores que poderiam ocorrer problemas, em territorio paraguaio, no rio Paraná a jusante de Itaipu, até a confluencia dos rios Iguaçu e Monday (R-11), devido as seguintes restriçoes:

· Casa de Força da Usína de Acaray
Nivel de Inundaçao = 128,50 m

· Vazao Máxima em R-11
51.000 m³/s (somatório das defluências de Itaipu e contribuiçoes do Iguaçu, Acaray e Monday)

A ELETROBRAS, através do CNOS, passou, entao, a exercer um continuo monitoramento das vazoes e níveis nas bacias dos rios Iguaçu e Paraná, em especial em R-11 e usina Acaray, em estreito e permanente contato entre as empresas ITAIPU, COPEL e ELETROSUL. Durante toda a noite do dia 29 e ao longo do dia 30. diversas manobras operativas foram efetuadas para salvaguardar as restriçoes, nas quais destacam-se:

· fechamento do vertedor da usina de Sao Simao, no rio Paranaíba;

· reduçao da geraçao da usina de Itumbiara;

· fechamento do vertedor da usina de Capivara, no rio Paranapanema;

· monitoramento continuo das linhas de transmissao da regiao Sul em funçao dos problemas decorrentes das condiçoes meteorológicas críticas;

· aumento das defluências em Itaipu;

· acompanhamento das vazoes e níveis nas seçoes ao longo do rio Iguaçu (Uniao da Vitória, Foz do Areia, Segredo, S. Santiago, S. Osório e Porto Capanema); do rio Paraná (Jupiá, Itaipu, usina Acaray) e dos principais afluentes a montante (Paranapanema, Capivara, Ivaí e Piquiri) e jusante (Monday = R-11);

Tais medidas tiveram como um dos objetivos principais a reduçao do armazenamento do reservatório de Itaipu, enquanto ocorria a propagaçao do pico da cheia na calha do rio Iguaçu. A medida em que a frente da cheia de aproximava da foz do rio Iguaçu, efetuar-se-ia a reduçao das defluências de Itaipu para respeitar a soma total das vazoes em R-11.

Com o monitoramento continuo das informaçoes nas diversas bacías foi possível reduzir as afluências em R-11, cujo valor máximo observado foi de 50.339 m3/s, as 3 horas do dia 31.05, resultante da liberaçao controlada de 20.939 m3/s em Itaipu somada as contribuiçoes incrementáis e á vazao do rio Iguaçu, o Anexo 6 ilustra a evoluçao das vazoes e níveis em Itaipu e nos rios Iguaçu, Paraná e Monday durante o período critico da cheia.

5.3 Consideraçoes Fináis

O problema do controle de cheias em um sistema de reservatórios é urna questao técnicamente complexa. Os bons resultados hoje colhidos so sao possíveis porque ao longo dos últimos quinze anos o setor elétrico brasileiro se preparou para atenderá esse desafío. Na maioria das empresas concessionárias, foram formadas e treinadas equipes técnicas especializadas em hidrología operacional, capazes de absorver os mais recentes avanços técnicos e desenvolver metodologias de controle de cheias voltadas para as características específicas do sistema brasileiro. Nesse sentido, cabe ressaltar o papel das equipes técnicas da ELETROBRAS e do CEPEL como propulsoras desse desenvolvimento.

Deve ser destacado o papel desempenhado pelo sistema de reservatórios da bacia do rio Paraná para a mitigaçao dos efeitos de inundaçoes. Como foi mostrado, as regras operativas dos aproveitamentos foram aperfeiçoadas visando nao apenas a geraçao de energía, mas também o controle de enchentes, acarretando beneficios socio-econômicos para toda a bacia, inclusive com reflexos favoráveis para os países ribeirinhos a jusante, como no caso do evento da cheia na bacia do rio Paraná, a montante de Jupiá, em janeiro/fevereiro de 1992 e do rio Iguaçu, em maio/junho do mesmo ano.

Julgamos importante citar algumas questoes técnicas que o setor elétrico precisará desenvolver e/ou aprimorar, pois, direta ou indiretamente, sao fundamentáis para os estudos de controle de cheias.

· Estudos hidrológicos para disponibilizaçao de series diarias de melhor qualidade;.
· Estudo de medidas para melhoria do controle hidráulico dos aproveitamentos hidroelétricos;
· Previsao meteorológica;
· Previsao de vazoes;
· Mapeamento das áreas inundáveis, com os riscos associados e sua ocupaçao socio-econômica.

Um outro aspecto importante a ser destacado é de que a implantaçao das medidas de prevençao de cheias acabam incentivando a ocupaçao de áreas inundáveis. Tendo em vista este fato, tem-se observado urna tendencia mundial de se dar um maior destaque as chamadas "medidas nao estruturais", tais como: regulamentaçao do uso do solo; seguro contra inundaçoes; previsoes de enchentes com sistemas de alerta; campanhas de esclarecimento ás populaçoes sobre as áreas sujeitas a inundaçoes.

No entanto, a proposiçao e implementaçao destas "medidas nao estruturais", que poderiam trazer um beneficio global para a sociedade, extrapolam o papel institucional do setor elétrico.

A necessária articulaçao entre os diferentes setores usuarios deverá ser facilitada num futuro próximo com a criaçao do Sistema Nacional de Gerenciamento de Recursos Hídricos - SINGREH, atualmente em discussao no âmbito do Congresso Nacional. A criaçao de um fórum amplo e específico para a discussao do uso múltiplo dos recursos hídricos permitirá urna análise abrangente dos aspectos técnicos e institucionais dessa questao.

A ELETROBRAS tem procurado estudar as questoes metodológicas do rateio de custos em obras de usos múltiplos e as técnicas de análise multi-objetivo em recursos hídricos. Essas atividades vem sendo desenvolvidas através de contratos com centros especializados em pesquisas em recursos hídricos, como por exemplo, o projeto de Usos Múltiplos de Recursos Hídricos realizado em conjunto com o Centro Tecnológico de Hidráulica da Universidade de Sao Paulo. Atualmente estao em andamento junto com o CEPEL - Centro de Pesquisas de Energía Elétrica um projeto para o aprimoramento das metodologias de controle de cheias, e outro relativo as consideraçoes de usos múltiplos ja na fase de inventario do potencial hidroelétrico de bacías hidrográficas.

Através do dominio das técnicas mais atualizadas e do perfeito equacionamento das questoes político-institucionais relacionadas ao uso múltiplo de recursos hídricos, o setor elétrico poderá cumprir com sua parte, assegurando a qualidade das medidas operativas de gerenciamento de seu sistema de resérvatenos e contribuindo para que ocorra um desenvolvimento económico e social do país em harmonia com o meio ambiente.

Referências bibliográficas

BARRETTO, L.A.L, LOPES, AL, GONTIJO, E.A, MING, L, KONISHI, S. (1979), "Regras de Operaçao de Resérvatenos para Controle de Cheias", V Seminario Nacional de Produçao e Transmissao de Energia Elétrica, Grupo I, GPH, Recife, PE, Brasil.

BEARD, L.R., (1963), "Flood Control Operation of Reservoir", Journal of the Hydraulics Division, ASCE, Volume 89, Proc. Paper 3380, pages 8-10, 21-23.

BEARD, L.R., (1977), Review of Flood Control Studies for Rio Grande, relatório de consultario para a ELETROBRAS, Rio de Janeiro, dezembro.

CECCA, (1977), Determinaçao de Volumes para Controle de Cheias nos Resérvatenos da Bacia do Rio Grande, Subcomitêde Estudos Energéticos, Grupo Coordenador para a operaçao Interligada, ELETROBRAS, Rio de Janeiro.

DAMAZIO, J.M, (1989), Condiçoes de Controlabilidade de Sistemas de Resérvatenos para Controle de Cheias e seu Uso na operaçao com Múltiplos Usos, Relatório Técnico, CEPEL N^o 036/89, Rio de Janeiro.

ELETROBRAS, (1987), Guia para o Cálculo de Cheia de Projeto de Vertedor, Rio de Janeiro.

GTEH, (1986), Modelo Cheia - Manual de Apresentaçao do Modelo de Simulaçao da operaçao Hidráulica para o Controle de Cheias em um Sistema de Resérvatenos, Subcomitê de Estudos Energéticos, Grupo Coordenador para operaçao Interligada, ELETROBRAS, Rio de Janeiro.

GTHO, (1995), III Encontró Técnico de Hidrología Operacional - Relatório Final, Subcomité de Estudos Energéticos, Grupo Coordenador para operaçao Interligada, ELETROBRAS, Rio de Janeiro.

GTHO, (1995), Prevençao de Cheias nos Aproveitamentos Hidrelétricos dos Sistemas Interligados Brasileiros - 1995/1996, Subcomité de Estudos Energéticos, Grupo Coordenador para operaçao Interligada, ELETROBRAS, Rio de Janeiro.

KELMAN, J., (1986), Stochastic Modeling of Hydrologic Intermitent Processes, Colorado State University Fort Collins, Colorado Hydrology Paper N^o 89, 1976.

KELMAN, J., (1987), Cheias e Aproveitamentos Hidrelétricos, ABRH, RBE, Rio de Janeiro.

KITE, G.W., (1977), Frequency and Risk Analyses in Hidrology, Fort Collins, Colorado.

MARIEN, J.L., (1984), "Controllability Conditions for Reservoir Flood Control Systems With Applications", Walter Resources Research, vol 20(11).

MING, L, BARRETTO, L.A.L, LOPES A.L, GONTIJO, E.A, KONISHI, S, (1979), "Prevençao de Cheias em Reservatórios do Sistema Interligado", V Seminario Nacional de Produçao e Transmissao de Energía Elétrica, Grupo I, GPH, Recife, PE, Brasil.

NERC, (1975), Flood Studies Report, Natural Environment Research Council, London.

ROCHA, V.F, DAMAZIO, J.M., KELMAN, J.,(1993), Modelo de Programado da operaçao de Controle de Cheias com Utilizaçao das Condiçoes de Controlabilidade, X Simpósio Brasileiro de Recursos Hídricos, Gramado, RS, Brasil

SIESE, (1994), Boletím Trimestral - Síntese 1994, Sistema de Informaçoes Empresariais do Setor de Energía Elétrica, MME - DNAEE - ELETROBRAS.

U.S. ARMY, (1959), Reservoir Regulation, U.S. Army, Corps of Engineers, May, pages 13-15.

USWRC, (1977), Guidelines for Determining Flood Flow Frequency, United States Water Resources Council, Washington.

VIEIRA. A.M, SALES, P.R.H, BARRETTO, L.A.L., (1986). "The Brazilian Eletric Sector Experience in Flood Control", Proc. of the Internation Symposium on Flood Frequency and Risk Analyses, Lousiana States University, Baton Rouge, U.S.A., pages 409-418.

Anexo 1 - Diagrama de Operação em Emergência - Reservatório de Furnas

Anexo 2 - Localização das Bacias do Rio Paraná e Iguaçú no Brasil

Anexo 3 - Diagrama Esquemático dos Aproveitamentos Localizados na Bacia do rio Paraná

Anexo 4 - Operação em Itumbiara e Fumas no período 01/01/92 a 20/02/92 (Reservatorio de Furnas)

Anexo 4 - Operação em Itumbiara e Fumas no período 01/01/92 a 20/02/92 (Reservatorio de Itumbiara)

Anexo 5 - Operação do Sistema de Reservatórios até Jupiá no período 01/01/92 a 20/02/92

Anexo 6 - Operação Hidráulica do Reservatório de Itaipu e Vazões Observadas na Confluência Paraná-Iguaçuno período 29/05/92 a 01/06/92

Anexo 4.3/Appendix 4.3 - The Mississippi basin flood of 1993

by Gerald E. Galloway, Jr. P.E., Ph.D.

EXECUTIVE SUMMARY

The Midwest Flood of 1993 was a significant hydrometeorological event. In some areas it represented an unusual event; in most others, however, it was just another of the many that have been seen before and will be seen again. Flood flows similar to those experienced by most of the Midwest can occur at any time. Flood damages are a national problem.

Excessive rainfall, which produced standing water, saturated soils, and overland flow, caused major damages to upland agriculture and some communities. In turn, runoff from this rainfall created, throughout the basin, flood events that became a part of the nation's 1993 TV experience. Damages overall were extensive: between $12 billion and $16 billion that can be counted, and a large amount in unquantifiable impacts on the health and well-being of the population of the Midwest. Over half of the damages sustained were agricultural damages to crops, livestock, fields, levees, farm buildings, and equipment. The remaining damages were primarily to residences, businesses, public facilities, and transportation. Much of the agricultural damage occurred in upland areas as the result of wet fields and a short growing season rather than inundation by floodwaters. Similarly a portion of residential and business damages was caused by basement flooding due to high groundwater and sewer back-up in areas outside the floodplain.

Rivers and river valleys historically have been major transportation routes, particularly in the area impacted by the 1993 flood. In the Midwest, transcontinental railroads, interstate highways, and other road systems either follow river valleys or cross them. As a result, physical damages to transportation systems created a significant percentage of total flood damages. In addition to direct damages, indirect costs accrued when transportation routes were inundated by floodwaters, and traffic was halted or detoured.

Human activities in the floodplains of the Midwest over the last three centuries placed people and property at risk. Local and federal flood damage reduction projects were constructed to minimize the annual risk, and, during the 1993 flood, prevented nearly $20 billion in damages. Some of these programs, however, attracted people to high risk areas and created greater exposure to future damages. In addition, flood control, navigation, and agricultural activities reduced available floodplain habitat and compromised natural functions upon which fish and wildlife rely.

Over the last 30 years the nation has relied heavily in the Midwest on a strategy of construction of levees and detention reservoirs to protect those at risk. It has, however, earned that multi-approach floodplain management can reduce vulnerability to damages and create a balance among natural and human uses of floodplains and their related watersheds to meet both social and environmental goals. The nation, however, has not taken full advantage of this knowledge. The United States simply has lacked the focus and the incentive to engage itself seriously in floodplain management. The 1993 flood focused attention on the floodplain.

A White House based Interagency Floodplain Management Review Committee investigated the causes of the 1993 flood and the nations program for floodplain management. It proposed that all levels of government, all businesses and all citizens should have a stake in properly managing the floodplain, and that those who support risky behavior, either directly or indirectly, must share in floodplain management and in the costs of reducing that risk. It noted that he federal government can lead by example; but state and local governments must manage their own floodplains. Individual citizens must adjust their actions to the risk they face and bear a greater share of the economic costs.

The Review Committee proposed a floodplain management strategy of, sequentially, avoiding inappropriate use of the floodplain, minimizing vulnerability to damage through both structural and nonstructural means, and mitigating flood damages when they do occur. By controlling runoff, managing ecosystems for all their benefits, planning the use of the land and identifying those areas at risk, many hazards can be avoided. Where the risk cannot be avoided, damage minimization approaches, such as elevation and relocation of buildings or construction of reservoirs or flood protection structures, are used only when they can be integrated into a systems approach to flood damage reduction in the basin. When floods occur, impacts on individuals and communities can be mitigated with a flood insurance program that is funded by those who are protected.

To ensure a long-term, nationwide approach to floodplain management, the Review Committee proposes legislation to develop and fund a national Floodplain Management Program with principal responsibility and accountability at the state level. It also proposed issuance of a Presidential Executive Order requiring federal agencies to follow floodplain management principles in the execution of their programs.

The upper Mississippi River Basin included both individually authorized federal flood damage reduction projects and levees built by local groups and individuals. This pattern of development is unique and requires a unique approach. The Review Committee proposed a plan for integrated management of the water resources of the basin.

INTRODUCTION

Throughout the Summer and Fall of 1993, the people throughout the world with access to the Cable News [Television] Network (CNN) were shown scenes of the devastation wrought on the Midwestern part of the United States by what became known as the Great Mississippi Flood of 1993. Since the early days of the 20th century, the Nation had labored to reduce the vulnerability of its people to flood damages and yet this flood had destroyed tens of thousands of homes, flooded hundreds of thousands of acres of prime farmland and had disrupted the economic and social fabric of several million people. National leaders as well as private citizens not only raised questions about how such flood damages occurred, but demanded to know what should be done to prevent recurrences of these damages.

This paper discusses the 1993 flood in terms of its extent and its impacts on the effected region and its populace. It is based on and extracts from the report a White House based Interagency Floodplain Management Review Committee that investigated the causes of the 1993 flood and the management of floodplains both in the Mississippi basin and nationwide. ¹

1. The Flood of 1993

1.1 The Upper Mississippi River Basin

The Mississippi and its tributaries have played a major role in the United States history. Their existence was critical to the growth of the Midwest region and fostered the development of major cities and a transportation network linking the region to the rest of the world. The floodplains of these rivers provided some of the most productive farmland in the country. They today offer diverse recreational opportunities and contain important ecological systems.

The Mississippi River basin stretches from the eastern slopes of the Rocky mountains in Montana to the western reaches of New York state (Figure 1). It drains all or parts of 32 states and 2 Canadian provinces covering 1.25 million square miles or 41 percent of the contiguous land area of the United States. The upper Mississippi River Basin (Figure 2) begins at the confluence of the Ohio and Mississippi Rivers and encompasses 714,00 square miles in 9 states, over 57 percent of the total basin. From its source at Lake Itasca, Minnesota, the Mississippi River courses a distance of 1366 miles. Its principal tributary, the Missouri River runs for over 2466 miles and drains 529,00 square miles above its mouth near St. Louis, Missouri. While the upper Mississippi flow is essentially unregulated except for the minimal controls need to operate locks and dams for navigation, flow on the Missouri is controlled by six large dams on the upper section of the river.

Land use in the region is predominantly for agriculture although there are clusters of industry in and near major cities and adjacent to many smaller communities. The areas 208 million cropland acres represent nearly 32 percent of United States farm acreage. In the floodplain of the upper Mississippi and Missouri, the same development patterns prevail with over 60 percent of the land (7 million acres) devoted to agriculture. Most major population centers in the region (except for Chicago) are located along the Mississippi and Missouri Rivers and their tributaries. Flood vulnerable sections of major urban areas, for the most part, have been protected by levees and/or reductions in flood stages brought about by upstream impoundments.

Figure 1. - The Mississippi Basin Source: Tobin and Montz, The Great Midwestern Floods of 1993.

Figure 2. - The Upper Mississippi River Basin

1.2 A Significant Hydrometeorological Event

The flood of 1993 was a hydrometeorological event without precedent in modem US history. The late summer and Fall of 1992 were wetter than normal for the Midwest and coupled with cooler temperatures produced wet soil conditions. A normal winter rainfall was followed by increasingly heavy precipitation throughout the spring, summer and into the early fall. Seasonal rainfall records were shattered in all nine states impacted by the flood. Summer rainfall amounts equaled those computed for storm frequencies having 75 to 300 year recurrence intervals. During the 14 month period from July 1992 to September 1993, rainfall amounts were significantly above average in all but 3 of the 14 months. In Iowa, the rainfall from January through September was the greatest amount, 44.5 inches, in 121 years of record. Evaporation was the lowest on record and cloud cover and soil moisture readings were the greatest in history.

This intense and continuous rainfall, coupled with the wet soil conditions, after a point in time, filled every ditch and channel, every stream and provided record flows on many reaches of the Missouri and Mississippi and their tributaries. By mid-July, river stages exceeded the 100 year discharge at 45 of nearly 500 gauging stations in the region and near 500 year elevations were seen along the Missouri from southeastern Nebraska to near St. Louis and from southern Iowa to above St. Louis on the Mississippi (figure 3)². The duration of the flood added to its impact. The high stages in July were followed in many reaches by late August and early September rises that approached the mid-summer levels. Considerable acreage was underwater for several months as continuing high stages prevented the drainage of floodplains.

Figure 3. - River Reaches with Significant Flooding (Source: US Department of Commerce, NOAA, National Weather Service)

1.3 Soil Conditions Prior to the 1993 Flood

The antecedent conditions that gave rise to the Flood of 1993 included, in addition to record rainfalls, wet soil conditions that began in the central Great Plains during the summer of 1992 and rose rapidly with the increasing precipitation and cooling air temperatures of late 1992. July, September, and especially November 1992 were much wetter than normal over the upper Mississippi River Basin. That winter precipitation was near normal, but a wet spring followed. By late March, extremely moist conditions covered much of the region as a result of the wet fall and spring snowmelt runoff. Iowa, which was centrally located in the area of heaviest flooding, experienced the second wettest November - April period in 121 years of record. This period was followed by above-normal precipitation over the upper Mississippi River Basin during April and May. The April - June period was the wettest observed in the upper Mississippi River Basin in the last 99 years. Consequently even before the onset of the heavy summer rains, most upper basin soils were saturated, and many streams and rivers were flowing at well above seasonal normal levels.

1.4 Rainfall

During much of the summer of 1993, a persistent atmospheric pattern of excessive rainfall occurred across much of the upper Mississippi River Basin. The major river flooding resulted primarily from numerous series of heavy rainfall events from June through late July. The recurrence of heavy rainfall was the direct result of a stable upper-level atmospheric circulation pattern with a deep trough to the west of the upper Mississippi valley and a strong ridge along the East Coast (Figure 4). In late July and early August, a change in the upper air circulation pattern brought drier conditions to the Midwest as the trough shifted eastward. Locally heavy thunderstorms generated some additional flooding in parts of the soaked upper Mississippi River Basin during mid-August; however, these rains were associated with a typical summertime pattern and not a return to the anomalous and persistent June and July atmospheric conditions.

Figure 4. - Weather Pattern, June-July 1993 (Source: US Department of Commerce, NOAA, National Weather Service)

During the June-August 1993 period, rainfall totals surpassed 12 inches across the eastern Dakotas, southern Minnesota, eastern Nebraska, and most of Wisconsin, Kansas, Iowa, Missouri, Illinois, and Indiana. Over 24 inches of rain fell on central and northeastern Kansas, northern and central Missouri, most of Iowa, southern Minnesota, and southeastern Nebraska. Up to 38.4 inches fell in east-central Iowa. Generally precipitation amounts were 200 to 350 percent of normal from the northern plains southeastward into the central Corn Belt.

Rainfall amounts over the upper Mississippi River Basin during the May-August 1993 period are unmatched in the historical records of the central United States. In July, broad areas in the lower Missouri River Basin experienced rainfall amounting to four times normal. The series of storms producing these record rainfalls were remarkable not only in their magnitude but also for their broad regional extent; record wetness existed over 26,000 sq. mi. of the upper Mississippi River Basin. Seasonal rainfall records were shattered in all nine states impacted by the deluge of 1993. Summer rainfall amounts equaled those computed for storm frequencies having 75-year to 300-year recurrence intervals.

1.5 River Row

The deluge across the upper Mississippi River Basin produced record setting peak flowrates and water levels in many tributaries and in the main stem rivers, including a large reach of the upper Mississippi, over the full reach of the middle Mississippi, and over much of the length of the lower Missouri River. Flooding began in the northern portion of the upper Mississippi River Basin in June and then moved southward with the shifting of the storm-producing weather pattern and the travel of the flood flows downstream as summer progressed.

Rainfall was particularly heavy between June 17 and 20 in southwest Minnesota and northwest Iowa, causing record flooding on the Minnesota River. The next major pulse of precipitation occurred from June 23-25. Runoff from these rains combined with flood flows from the Minnesota River to initiate the first flood crest that moved down the upper Mississippi River. Following a short, dry period, a prolonged siege of heavy precipitation occurred from June 30 to July 11. This included extreme amounts of rainfall on July 9 in Iowa, which produced record flooding on the Raccoon and Des Moines rivers. Just as the crests from these two rivers reached Des Moines, a relatively small, connective pocket dumped several inches of rain on the crests rapidly boosting the river levels and flooding the city's water treatment plant. The intense rainfall during this period also led to record flooding on portions of the lower Missouri River and combined with the crest already rolling down the Mississippi to establish record river stages from the Quad Cities area on the upper Mississippi River downstream to Thebes, Illinois, on the middle Mississippi River. Another major precipitation event occurred from July 21-25. The heaviest rains were focused farther south than the earlier events, with especially heavy rain falling over eastern Nebraska and Kansas, leading to the second major crests on both the Missouri and Mississippi rivers.

Kansas City and St. Louis saw two flood peaks, one caused by the June 30 to July 11 rainfall and the other by rain falling from July 21-25. While flooding from the latter rainfall period did not extend as far upstream on the Mississippi River, new record river levels occurred at many locations downstream and on much of that portion of the Missouri River that flows through Missouri. Above normal rains continued to occur over parts of the flood-affected region during August, especially over Iowa where accumulations were twice the normal monthly amount over much of the state. By mid-September, however, rainfall began to diminish and rivers began to recede. Then, at the end of September, a strong system of thunderstorms deposited 1 to 3 inches of rain over the State of Missouri and 7 inches or more from the central part of the state eastward. The consequence was major flash flooding on many tributaries and new flood crests on the lower Missouri and middle Mississippi rivers. Farmlands behind previously breached levees were reflooded and two people drowned in separate incidents. Many roads were washed out and there was much damage to property in Missouri.

2. Impacts of the 1993 Flood

2.1 General

The costs to the nation from the flood were extensive. Thirty-eight deaths were attributed directly to the flood and estimates of fiscal damages ranged from $12 billion to $16 billion. Agriculture accounted for over half of these damages. More than 100,000 homes were damaged. Flood response and recovery operations cost more than $6 billion. Many costs could not be quantified as impacts on businesses in and out of the basin were difficult to calculate and there was no accurate way to assess tax losses to governments. There also were impacts of the flood on the population's physical and mental well-being, but these too were difficult to sum.

The Midwest Flood of 1993, one of the most costly flood events in this nation's history, flooded over 6.6 million acres in the 419 counties in the upper Mississippi Basin. The damages experienced reflected the land-use and settlement patterns within and adjacent to the floodplain. The floodplains along the main stem Mississippi and Missouri rivers and the major tributaries that were inundated generally are used for agriculture, and most areas are sparsely populated. Throughout most of the area, river towns are protected by urban levees, or they are located primarily on a bluff. Floodwaters thus inundated neighborhoods rather than entire communities. Residences, businesses, and industries did receive extensive damages in bottomland areas and along tributaries near Kansas City and St. Louis. Development in these urban areas, however, is largely in the uplands or protected by urban levees that provided flood protection. As a point of comparison, significantly fewer people were impacted by the Midwest Flood of 1993 than were impacted by the 1927 flood on the lower Mississippi River.

Over half of the damages sustained were agricultural damages to crops, livestock, fields, levees, farm buildings, and equipment. The remaining damages were primarily to residences, businesses, public facilities, and transportation. Much of the agricultural damage occurred in upland areas as the result of wet fields and a short growing season rather than inundation by floodwaters. Similarly a portion of residential and business damages was caused by basement flooding due to high groundwater and sewer back-up in areas outside the floodplain.

The National Weather Service (NWS) estimated damages for the Midwest flood at $15.7 billion based on information provided by its field offices. This estimate was based on totals by state, but did not include breakdowns of damage by type. In August 1993, The New York Times published an estimate of nearly $12 billion in damages based on information it obtained from state and federal officials. State and federal officials could not assess all damages until floodwaters receded, and the full extent of agricultural damages was not known until after the end of the growing season. Most of the affected states updated their damage estimates, and the total ranges from $12 billion to $13 billion.

The Review Committee developed an estimate of flood damages using federal payments and making assumptions as to what percentage of damages those payments represent. This information indicates that total damages were more than $12 billion with as much as $4 billion to $5 billion of that total being agricultural damages in upland areas.

Damage estimates for the Midwest flood show marked inconsistencies. No federal agency was responsible for developing accurate assessments of flood damages, nor was funded to do so. The affected states and the Federal Emergency Management Agency (FEMA) conduct preliminary damage assessments to determine if a Presidential disaster declaration is warranted and to estimate the resources necessary for response and recovery. Once sufficient damage has been identified to justify a declaration and once FEMA has a general idea of how resources should be allocated, federal agencies have little incentive to expend resources updating preliminary assessments. Resources are instead focused on tracking and projecting expenditures. The NWS is not funded to estimate total damages but does so to support other missions. The US Army Corps of Engineers (USACE), which in the past estimated flood damages, is no longer funded to do so. (The Review Committee was concerned that decisions involving hundreds of millions of dollars often were made without systematic assessments of flood damages and without a clear understanding by decision makers of the nature and extent of the damages incurred).

2.2 Agriculture

Agriculture is the leading industry in most counties of the nine states affected by the Flood of 1993). The area's 208 million cropland acres represent 32 percent of the nation's farm acreage, 35 percent of total agricultural sales, and almost 60 percent of total national com, wheat, and soybean acreage. Combined production from Illinois and Iowa alone represent 33 percent of com and 30 percent of soybean acreage in the United States, but dominant crops and yields vary by state throughout the region. Floodplains comprise approximately 11 percent of total acreage affected by the 1993 flood and 66 percent of this acreage is in agricultural production.

Agricultural damages from the Flood of 1993 had two primary causes: excessive moisture that prevented planting and reduced yields in upland and floodplain areas and actual flooding that destroyed crops and severely damaged many acres of fertile floodplain cropland. It is difficult to separate the factors that influenced crop production during the 1993 growing season in the 9-state region. They included rain, low temperatures, early frost, and floods. More than 70 percent of the crop disaster assistance payments, however, were made to counties in upland areas - not in main stem river floodplains.

Agricultural damages directly attributed to actual flooding totaled more than $2.5 billion, with an estimated $1.4 billion in lost com and soybean sales. Most of these losses were restricted to 1993 as the productive capacity of the land was unchanged. There were, however, damages to field fertility and farm infrastructure of at least $100 million.

Each state suffered different types of losses. For example, Missouri with 34 percent of its cropland (5.1 million acres) in the floodplain, had crop damages from flooding on 3.1 million acres causing $247 million in lost sales (Cassidy and Rickert, 1994). In Illinois, only 3 percent of the state's com and soybean acreage (312,000 and 276,000 acres respectively) were lost to flooding with a loss in sales of $153.4 million (Bhowmik, 1993). Minnesota farmers lost $500 million in crop sales, but most of the damage was caused by wet conditions rather than riverine flooding (Taff and Maki, 1993). Damage from scour and deposition affected 455,000 acres on the Missouri River floodplain representing 20 percent of the flooded cropland along the Missouri and Mississippi rivers. Drainage ditches were filled with sediments, and other agricultural infrastructure was destroyed. Almost 60,000 acres have sand deposition more than 24 inches thick and reclamation costs to restore fertility to damaged cropland were estimated at $190/acre. If cropland restoration required removal of sand, it cost approximately $3,200 to remove each acre-foot of sand. Over $10 million was required remove sediment and debris from ditches (US Department of Agriculture, 1993). Secondary impacts of agricultural losses to a local economy vary substantially with the dependence of that economy on the agricultural sector. Immediate losses are due to lost sales and unemployment. In the long run, the assessed value of land that sustained long-term damage may be reduced which will affect the property tax base of affected communities.

Another secondary effect was a reduction in crop-support payments since crop prices adjusted to the reduced production caused by wet weather in the Midwest and drought in the Southeast in 1993. This loss to farmers was a gain for taxpayers since subsidies represent transfer payments. For com, these deficiency payments were reduced by more than $2.6 billion (US Department of Agriculture, 1993). These price effects and subsequent reduction in deficiency payments were temporary, as the 1994 crop supply returned to past levels.

It is important to note that the majority of 1993 agricultural damages in the Midwest were caused by wet soil conditions and inundation in upland areas. Damage to inundated cropland in the floodplain was significant with almost complete crop losses behind failed levees. Areas affected by severe erosion and deposition may suffer long-term loss of productivity.

2.3 Residences And Businesses

Estimates vary on the number of homes flooded and families impacted by the Midwest flood. Surveys made by Red Cross workers immediately after the floods identified more than 55,000 flooded residences (American Red Cross 1993). FEMA subsequently verified these damages with Red Cross chapters and developed an updated estimate of 70,545 residences (Shepard, 1994). The New York Times (August 10, 1993), estimated that more than 84,000 residences were damaged. By April 1994, the federal government had received 167,224 registrations for individual assistance and 112,042 applications for the Disaster Housing Program.

Within this latter group, over 90,000 applications were approved. The Disaster Housing Program data indicated that more than 100,000 residences were flooded (US FEMA, 1994).

Businesses sustained significant physical damages particularly in urban areas such as St. Louis County and the Kansas City areas of Missouri. Much of this damage occurred behind levees that failed or were overtopped. The 996 National Flood Insurance Program (NFIP) claims payments made to small businesses) and the 4,667 Small Business Administration (SBA) loans for damages to businesses indicated that in excess of 5,000 individual businesses were damaged (US FEMA, FIA, 1994). No overall damage estimates for businesses were available, but as a measure of this damage, SBA loans to businesses, exceeded $334 million for physical damage and economic injury (Kulik, 1994). Add to these loans NFIP flood insurance payments for small businesses and other non-residential buildings that exceeded $94 million, and the total exceeds $431 million (US FEMA, 1994). In addition to physical damage to buildings and their contents, lost profits and wages from businesses closed by the flood had local and regional impacts. For example, an American Cyanamid Plant near Hannibal, Missouri, was protected by its own levee and not damaged by floodwaters, but the plant was shut down for nearly three months because its access road was inundated when an agricultural levee failed.

2.4 Transportation Systems

Rivers and river valleys historically have been major transportation routes, particularly in the area impacted by the 1993 flood. In the Midwest, transcontinental railroads, interstate highways, and other road systems either follow river valleys or cross them. As a result, physical damages to transportation systems created a significant percentage of total flood damages. In addition to direct damages, indirect costs accrued when transportation routes were inundated by floodwaters, and traffic was halted or detoured.

A major portion of flood damages to public facilities in 1993 involved roads and bridges. These damages ranged from blown culverts and wash-outs on rural roads and city streets to loss of bridges and damages to interstate highways inundated by floodwaters. The repair of flood-damaged roads and bridges generally is funded through the FEMA Public Assistance Program or the Department of Transportation. Funds expended by those agencies when added to the state/local cost share for public assistance indicated that total physical damages to roads and bridges exceeded $250 million (US FEMA, 1994).

· Roads and Bridges

Road and bridge flooding caused indirect losses related to increased transportation costs. In extreme cases, detours of 100 miles were required to travel between adjoining communities that had been connected by a bridge. Often bridges were elevated high above the river to allow for navigation or to minimize hydraulic impacts of floods, but bridge approaches built at or near the natural elevation of the floodplain were inundated by floodwaters. Even though the bridge was undamaged and the approach damage was minimal, the economic impacts on the communities served by the bridge could be extreme, particularly for a long duration flood such as occurred in 1993. For example, Keokuk, Iowa, was cut off from market areas in Illinois and Missouri for several weeks when the approaches to bridges over the Mississippi and Des Moines rivers were inundated. This resulted in serious economic impacts on local businesses. Flooding of the approaches to the bridge over the Mississippi River at Quincy, Illinois for 73 days resulted in an estimated $30 million in lost business to Quincy merchants (Bhowmik, 1993). In addition, many people who lived in Missouri and could not commute to work in Illinois were temporarily unemployed. Ferries were eventually established to address part of this problem. The full magnitude of these losses were reflected in over 36,000 claims approved for a total of $92 million in Disaster Unemployment Assistance.

· Railroads

Historically railroads were built in floodplains and river valleys to minimize construction and fuel costs. Main lines continue to parallel both the Missouri and Mississippi rivers. Although generally tracks are elevated on embankments above the elevation of most floods or are located behind levees, they remain subject to major flood events. In 1993 over 800 miles of track were flooded and several main lines were inundated for varying periods of time, but most trains were routed around flooded areas. The Association of American Railroads estimated that railroad damages totaled $182 million, including $131 million in physical damages to tracks, bridges, signals, communication lines, switches, locomotives, rolling stock, and buildings. Additional costs of $51 million resulted from detouring trains around sections of flooded track (Harper, 1993). Repair costs were generally borne by the railroads themselves although $21 million was distributed to railroads through a supplemental congressional appropriation: "Local Rail Freight Assistance."

· Airports

Airports often are located in floodplains because of the flat terrain and close proximity to urban areas. The Federal Aviation Administration (FAA) identified 33 airports with varying degrees of flood damage and repair costs exceeded $5.4 million. The airports ranged in size from the Spirit of St. Louis Airport in St. Louis County, Missouri, to airports that were little more than grass landing strips with a few hangars for private aviation. Most of the flooded airports were in Missouri (16) and Iowa (12). The Spirit of St. Louis Airport, an alternate for Lambert-Airport, sustained $1.7 million in damages when a 100-year local levee failed. Other major airports that were flooded include those at Creve Couer and Jefferson City and the Kansas City Downtown Airport. Several smaller airports went out of business (Trilling, 1994).

· Navigation

Two types of navigation projects are present in the basin. One, on the upper Mississippi and Illinois Rivers, is slack water navigation created and controlled by a system of locks and dams. The other, open water navigation, is utilized on the Missouri River and middle Mississippi River (the Mississippi below St. Louis). The upper Mississippi River navigation system provides a variety of uses: commercial transportation, recreation, environmental resources, water supplies for domestic and industrial use, and energy production. The Water Resource Development Act of 1986 declared the upper Mississippi River system to be a nationally significant ecosystem and a nationally significant commercial navigation system.

Navigation on the Mississippi River was a primary factor in settlement of the valley. The federal government began to support commercial navigation actively in 1824; first with 4-foot deep channels, then 4.5-foot and then 6-foot channels. The navigation channel projects, authorized by Congress in the 1930s for the Mississippi and Illinois Rivers, extended 9-foot draft navigation upstream to Minneapolis/St. Paul and connected the St. Lawrence-Great Lakes with the Mississippi-Ohio-Missouri navigation systems (Figure 5). The upper Mississippi River 9-foot navigation project has converted the Mississippi River (St. Louis to Minneapolis/St. Paul) into a series of pools at low and normal flow. Navigation dams, each consisting of a row of gates mounted between piers over a low sill, are used to maintain sufficient water depth for navigation. During periods of high flow, the navigation gates are completely opened to allow passage of the flood flows. Construction of the 29 lock and dam projects on the Mississippi River north of St. Louis was completed by 1950. These locks are nearing the end of their economic life span and may soon start to require expensive replacement. Locks and Dam 26 near, Alton Illinois, was replaced during the early 1980s at a cost of nearly $1 billion. Below the southern most lock, Lock 27 at Granite City, Illinois, navigation is maintained through placement of flow regulating structures such as wing dikes and by dredging that channelize, narrow, and deepen the river.

Figure 5. - Upper Mississippi River Navigation System (Source: Upper Mississippi River Basin Commission, 1982)

Maintenance of the upper Mississippi River navigation system requires periodic dredging at over 200 sites, removing an average of 9.5 million cubic yards of material annually. Two construction projects have supported navigation activities on the Illinois River. The first, the Chicago Sanitary and Ship Canal, completed in 1900, diverted water from Lake Michigan into the Illinois River. The second, a modem lock and dam system, similar to that in operation on the upper Mississippi River, consists of seven separate navigation locks. This system was completed in 1965.

In 1945 Congress authorized a comprehensive navigation plan for the Missouri River system. The result was a 9-foot channel navigation project to channelize and deepen the river from St. Louis upstream to Sioux City, Iowa. Six multi-purpose main stem reservoirs, affecting over 900 river miles, were developed above Gavins Point Dam. One purpose was to provide a regulated release of water for downstream navigation. Downstream of Gavins Point Dam, the river consists largely of a 735-mile navigation channel maintained with wing dikes, channel stabilization and other erosion and sedimentation control devices. Annual water release for navigation is based upon available water supplies. Navigation needs combined with winter releases for water supply and hydropower demands obligate all available water during a normal year. The navigation season on the Missouri River is limited to the ice-free season between April 1 and December 1.

Most of the Upper Mississippi, Missouri and Illinois Rivers were closed to barge traffic from July 11 until August 15, 1993 since the locks were out of operation and movement along the open waterways created wave action that endangered the flood control levees. Severe limitations on barge traffic continued through September, October, and November. The Maritime Administration estimated that losses of revenue to the navigation industry were $300 million per month. More than $165 million were lost in Illinois alone. Regional impacts on jobs from barge and port disruptions were not quantified but were estimated to be greatest in Illinois (US Department of Transportation, 1993).

2.5 Public Facilities and Utilities

The Midwest flood caused extensive damages to water and wastewater treatment plants and other public facilities. Damages to utilities, including water and wastewater treatment facilities and stormsewer systems, exceeded $85 million. Water treatment plants often are located in floodplains to be near well fields or the surface water that supplies the system. In addition, water supply lines must cross floodplains to serve floodplain residents. The EPA identified 200 municipal water systems impacted to some degree by the flood. The most prominent example was the Des Moines Water Works that served the City of Des Moines and adjoining communities. The plant was flooded and remained out of operation for 12 days and water from it was not safe to drink for another seven days. In addition to physical damages of $12 million, significant impacts were felt in the service area. Businesses and government offices closed because of lack of fire protection, and bottled water and portable toilets had to be provided for residents. The economic impact of the shutdown exceeded the cost of repair of the physical damage.

Wastewater treatment plants also tend to be located in floodplains which are generally the lowest point in a community and offer the advantage of gravity flow. Furthermore the effluent from these plants is discharged into major rivers or streams. The impact of flooding ranges from temporary plant shutdown and the discharge of raw sewage into the river during the flood to physical damage that results in extended plant shutdowns and continued discharges of raw sewage or partially treated effluent until such time as the plant can be repaired. A total of 388 wastewater facilities were impacted by the flood (Knight, 1993). Damages to public buildings exceeded $27 million. Water control facilities had more than $20 million in damages, and facilities such as parks and other recreation facilities recorded more than $22 million. These estimates were based on FEMA projections of infrastructure spending that included a 10-percent local cost share.

In general, energy facilities were not directly affected by the flood. Careful siting of the facilities and construction by the utilities of flood control works protected most plants from inundation. However, the previously discussed transportation problems with roads, bridges, railroads and navigation did restrict the activities of many energy systems. Employees were not able to get to the plants and fossil fuel supplies by railroad and barge were disrupted. Although no plants were shut down as a result, several approached the end of their on-site supplies.

2.6 Environmental Hazards

The EPA determined that 59 Superfund sites experienced flooding; however, impacts to the sites were minimal and corrective measures were completed on sites requiring them. In addition 73 solid waste treatment, storage, and disposal sites were flooded. Large propane tanks that were dislodged, floated downriver and created the potential for massive explosions. Besides the large propane tanks, states collected over 18,000 orphaned drums - each with a potential hazardous or toxic substance - and a large amount of household hazardous wastes whose disposal was necessitated by the flooding (US EPA, 1994). Daily loads of agricultural chemicals (herbicides and nitrates) transported by the Mississippi River were large relative to previous years; record flooding did not dilute the concentrations of herbicides. Concentrations of two herbicides (atrazine and cyanazine) in some samples from the Mississippi River exceeded health-based limits for drinking water. (Goolsby and Battaglin, 1993). The cumulative impact of any flood-related releases of hazardous materials, including pesticides, herbicides, and other toxic materials has not been yet established. The effects of flooding on groundwater hydrology and groundwater quality also have yet to be determined. In response to concerns regarding the safety of private wells, the Administration established a well-water contamination survey in coordination with the nine flood states (Young, 1994). The EPA performed flood water quality sampling around major metropolitan areas on the Missouri River. In some cases, drinking water standards were exceeded, but the majority of the readings posed no health risk. Results from sampling of treated drinking water revealed three locations where maximum contaminant levels were exceeded although results from a single sample did not indicate a problem (US EPA, 1993). (USGS) and the National Oceanic and Atmospheric Administration (NOAA) did not find significant changes in water chemistry since the 1993 flood (Goolsby and Battaglin, 1993). Impacts of the flooding on the distribution of contaminated river sediments is also unknown. Studies are underway to determine sediment chemistry and characterize sediment deposition patterns in rivers and streams (US EPA, 1993).

Effects of the flood on public and mental health are largely anecdotal. Some communities noted increases in spousal and child abuse and numbers of calls for police response. Mental health effects of community and individual buyout/relocation are poorly understood. Several studies are underway to assess the human response to the 1993 flood and to evaluate the factors that strain the ability of families to function adaptively to the event. Experience with other floods indicates that outbreaks of Equine, Western, and St. Louis can be expected two years after a flooding event (due to the lag time in amplification of disease vectors). The length of time between the flood event and the appearance of disease adds to the problem of attributing costs (Young, 1994).

3. Vulnerability Reduction

3.1 General

The rivers and streams of the Midwest were focal points for early settlement because they provided sources of drinking water and avenues for transportation and trade. Once settlements were established along rivers, the problem of controlling floods to protect human life and investments became readily apparent. At first small mounds of dirt were thrown up to divert water away from towns, and over the course of time, these mounds became levees and floodwalls. Many people living in floodplains behind those levees and floodwalls remain at risk because of decisions made many years ago. The modern challenge is to reduce those risks.

As settlers spread west they altered prairie, forest, meandering streams, and free-flowing river landscapes to provide arable farmland, raw materials for homes and industry, and transportation. Federal policies encouraged extensive private land development which then required construction of reservoirs and levees for flood protection. Human use thus changed Midwestern landscapes to the detriment of natural ecological systems. The Flood of 1993 raised questions as to what extent these landscape changes contributed to flood frequency and duration.

3.2 Levees and Reservoirs

A flood in 1927 affected millions of people throughout the Mississippi River Basin and demonstrated the inadequacy of the pattern of private flood damage reduction measures begun in 1879. It became a milestone event leading to major changes in national floodplain management policy. The 1928 Flood Control Act, which established a lower Mississippi River flood damage reduction system, and the 1936 Flood Control Act were the first codification of the federal interest in the coordinated development and installation of flood damage reduction measures. The primary method used to prevent damages in those early years was floodplain levees. Starting in 1936 the USACE responsibilities were focused on major rivers and development of congressionally approved plans for reservoirs, levees, channelization, and diversions. The methods used were those determined to be most cost effective for preventing flood damages. The USACE has constructed 76 reservoirs in the upper Mississippi River Basin. These control a drainage area of almost 370,000 square miles and contain a total flood storage volume of 40 million acre-feet of water. Forty-nine are located in the Missouri River Basin where the USACE also operates 22 Bureau of Reclamation reservoirs for flood storage. The majority of the reservoirs are operated to provide benefits on the tributaries where they are located; some are operated to benefit the main stem rivers. In addition to the reservoirs, the USACE has constructed or improved over 2,200 miles of levees for the protection of communities and agriculture in the upper Mississippi River Basin. Though records on the federal levees are kept by the USACE), at the time of the flood, there was no inventory about the estimated 5800 miles of non-federal levees that are in the upper basin.

Flood damage reduction-related activities of the US Soil Conservation Service (SCS) began nationally in 1944 with passage of PL 78-534 authorizing installation of upland treatment and flood damage reduction works in selected watersheds. (In 1995, the SCS was retailed the Natural Resource Conservation Service.) The Watershed Protection and Flood Prevention Act of 1954 (PL 83-566) expanded the SCS flood damage reduction program to the entire nation. During the past 42 years, in the nine Midwestern states affected by the Flood of 1993, the SCS has planned and evaluated over 300 watershed projects covering 40,000 sq. mi. (25.5 million acres). Locally sponsored PL-566 projects resulted in the installation of 2,964 reservoirs that influence the drainage of over 5 million upland acres, and 818 miles of channel work, 75 percent of which is located in North Dakota, Minnesota, and Illinois. The SCS requires 75 percent of the land above a proposed reservoir site to be treated before construction. It is estimated that PL-566 resulted in soil and water conservation treatments on more than 3 million upland acres.

The 1993 flood demonstrated that dams and reservoirs, engineered and built to store and regulate floodwater discharge, can reduce flood damages. All federally funded flood storage reservoirs operated as planned during the 1993 flood. At some facilities, emergency spillway flows occurred when inflow volume exceeded reservoir storage capacity. During the period of peak flooding (April 1 to August 1, 1993), the USACE reservoirs stored 22.2 million acre-feet of flood water. Approximately 18.7 million acre-feet were stored in the Missouri Basin, half of which was stored in the 6 main stem Missouri River reservoirs. Most of the remainder was stored in tributary reservoirs of the Kansas and Osage rivers. About 3.5 million acre-feet of water was stored in the Mississippi River Basin and an additional 1.1 million acre-feet were stored in 2,964 small PL-566 upland flood damage reduction reservoirs. Flood damage reduction reservoirs effectively controlled excess runoff and reduced damages to downstream floodplains during the 1993 flood event. The combined effect of the storage of flood waters in the federal flood damage-reduction reservoirs in the Missouri River basin reduced the average discharge of the Missouri River near its mouth, during the month of July, by 211,000 cfs. This had the effect of lowering the peak stage of the Mississippi River at St. Louis by 5 feet.

Federally constructed levees, in concert with upstream flood-storage reservoirs, protect many large urban areas from potentially significant damage. For example, without levees or floodwalls, portions of low lying areas in Rock Island and Moline, Illinois, and Kansas City would have been devastated. At St. Louis the Mississippi River crested at 49.6 feet on the USGS gage, almost 20 feet above flood stage, yet that portion of the city protected by the large flood wall escaped inundation. During and following the flood there was considerable speculation about the effect of levees on flood levels. Most of the speculation was based upon inferences drawn from comparisons between recent event data, obtained from systematically-measured river flow (discharge) and river level (stage) records, and similar data for historical floods. Such discussions failed to recognize that significant differences in data quality exist between the modern (after 1930) and the historic record. In addition, many other changes have occurred in the upper Mississippi River Basin which have created differences in flow regimes over time.

To ascertain the actual effect existing levees had on peak 1993 Mississippi and Missouri river flood stages, the Review Committee's Scientific Assessment and Strategy Team (SAST) commissioned model runs to analyze unsteady state river flow conditions in several river reaches of the Mississippi and the Missouri. The analyses suggested that if all the levees (other than urban levees) were absent, the peak stage at St. Louis in 1993 would have been reduced by 2.5 feet, but still more than 17 feet above flood stage and almost 4 feet higher than the previous known maximum level recorded during the 1973 event. An independent model commissioned by the St. Louis Post-Dispatch showed that the overtopping and breaching of two levees downstream from St. Louis reduced peak stage at St. Louis by 1.6 feet.

Based on these analyses and previous studies of these rivers, the Review Committee concluded that levees did not cause the 1993 flood. During large events such as occurred in 1993, levees have minor overall effects on floodstage but may have significant localized effects.

Federal Flood damage reduction projects and floodplain management programs, where implemented, worked essentially as designed and significantly reduced the damages to population centers, agriculture, and industry. Reservoirs and levees built by the US Army Corps of Engineers (USACE), prevented more than $19 billion in potential damages. Large areas of Kansas City and St. Louis were spared the ravages of the flood, although several suburbs suffered heavy damages. Watershed projects built by the Soil Conservation Service saved an estimated additional $400 million. Land use controls required by the National Flood Insurance Program (NFIP) and state floodplain management programs also reduced the number of structures at risk throughout the basin although a specific dollar value can not be assigned to these loss prevention efforts. Although flood damage reduction reservoirs and levees reduce the risk of flooding, they do not eliminate it. Given enough rainfall the flood damage reduction storage capacity of a reservoir can be exceeded and water will overtop the spillway. Local flooding may then occur downstream; its extent will depend upon the condition of the stream when the overtopping occurs. Throughout the basin, the Flood of 1993 exceeded the design capacity of many levees and the flood storage capacity of some reservoirs, flooding lands and property of persons who may have thought they were not at risk.

(During the 1993 flood many people suggested that the flood crest could be lowered significantly by opening navigation dam gates before the arrival of flood waters. Hydraulic investigations by the University of Iowa and evaluations of the 1993 flood show that navigation dams cause slight, localized increases in flood height just upstream of a dam. They did not cause increases in flood elevations for the entire Mississippi River System. In the middle Mississippi (from St. Louis to the confluence with the Ohio River) and on the Missouri River, navigation channels have no locks and dams, and the dikes and revetments which are in place cause little or no restriction to flow. The Review Committee concluded that navigation dams and locks did not cause an increase in the stage heights of the 1993 flood.)

3.3 Wetlands and Upland Treatment As Vulnerability Reduction Mechanisms

Wetlands occur in poorly drained soils and in areas where water is found at or near the ground surface. Between 1780 and 1980 an estimated 53 percent of the nation's original 221 million acres of wetlands were drained. In the nine Midwestern states affected by the flood 57 percent of the wetlands have been converted to other uses. The Swamp Land Acts of 1849, 1850, and 1860 resulted in the transfer of nearly 65 million acres of wetlands in 15 states from federal to state administration for the purpose of expediting their drainage. United States policy from the mid to the late 1800s has been to cede "overflow and swampy" lands to the states and to convert these lands to productive use. Substantial bottomland timber harvesting began with the arrival of pioneers, and by the 1930s, most wetlands had been converted from natural to agricultural uses and over 84 million acres nationwide had been included in regional enterprises known as drainage districts. By the 1950s, forested wetlands had been reduced to 66.7 million acres, and by the mid-1970s an additional 6.5 million acres had disappeared. Between the mid-1950s and 1970s an average of 458,000 wetland acres were lost each year in the coterminous United States. Agricultural development was responsible for 87 percent of the loss as wetlands were drained, filled, or otherwise converted to cropland. Development in the upper Mississippi River Basin for agriculture and other economic activity, flood damage reduction and navigation greatly altered the original landscape. The characteristics of flood events and the modification of the basin's natural resources reflect these changes.

Upstream land use and land treatment affect downstream flow regimes of rivers and floodplains. In considering floods and floodplain management, knowledge of where and how runoff occurs and which land practices can hold the rain where it falls for as long as possible become critical. Proper management can greatly affect the quantity and quality of water and sediment transported by flood waters. Factors influencing the amount and velocity of runoff include the amount and intensity of precipitation, soil type, land slope, available storage and land cover. Proper management of agricultural lands requires use of protective cover or land conservation practices. In the Midwest cropland erosion can be reduced by using measures such as conservation tillage, terraces, crop rotations, field borders, sediment and debris basins, strip cropping, and permanent vegetation. Such land use practices increase infiltration rates and help hold both water and soil in place. It is estimated that 37 percent of the nation croplands have adequate land treatment installed.

Pre-1850 historical records indicate that even prior to the clearing of wetland areas major floods occurred in the Mississippi River Basin. As part of economic development in the Midwest a substantial percentage of agricultural lands were created by drainage of wetlands and hydric soils. Hydric soils, good indicators of past and present wetland locations, total 10.4 percent of Mississippi and Missouri basin soils. The Review Committee heard numerous times that flooding would have been reduced had more wetlands been available for rainfall and runoff storage. An evaluation of the upper Mississippi River Basin's capacity to store rainfall runoff estimates that the soil profile has 10 times more storage capacity than above ground storage in depressional potholes. Because much of the basin was wet in 1993, particularly the areas that received the highest rainfall, the buffering capacity of the basin was depleted and unable to store water from the rains of June and July.

Topography has a direct impact on water movement and soil formation. The upper Mississippi River Basin is characterized by two distinct kinds of landscape: open systems which drain externally, and closed systems where drainage is trapped within a common depository. Due to the extended period of rain preceding the 1993 flood, the impacted area became completely saturated and surface depressions filled; therefore, storage available for additional runoff could only be found in the deep depressions areas located in the prairie pothole region of the Dakotas, Minnesota and Iowa. Hydrologic model studies of four watersheds that are representative of distinctly different upper Mississippi River Basin areas or terrain units were completed in 1994 by the SAST. The modeled watersheds represented only 5 of the 70 terrain types in the basin and therefore information derived from these models has limited applicability to assessing flood flow reductions basin-wide. In the watersheds modeled the maximum reduction for floodplain wetlands was 6 percent of the peak discharge for the 1-year event and 3 percent of a 25- and 100-year storm event. Wetlands are more effective in upland areas with more deeply incised potholes, where reductions were 23 percent of the 1-year event, 11 percent of the 25-year event, and 10 percent of the 100-year event. In areas of shallow depressions, restored wetlands reduced peak discharge by 9 percent of the 1-year event, 7 percent of the 25-year event, and 5 percent of the 100-year event. With the installation of a combination of land treatment measures and restored wetlands in the watershed, the models indicated runoff reductions of 12 to 18 percent are possible for the 25-year or less event. This indicates these practices could be effective for the smaller storm events. Wetland restorations in the uplands could function much the same as small upland reservoirs. It was shown more than three decades ago that small flood damage reduction dams are effective in the reach of stream immediately downstream but their effect diminishes rapidly with distance. Another report concluded that for certain watersheds, peak flow decreases as wetland areas increase. In very small watersheds (less than 100 sq. mi.), peak flowerets decreased by an average of 3.7 percent for each increase in wetland area equivalent to one percent of the area of the watershed. Applicability of this report may be limited only to the study areas. While wetlands may have some impact on peak flow in the smaller watersheds during smaller storms, their effects in larger watersheds during larger events has not been sufficiently documented and needs further study. Alternative watershed practices clearly produce varying degrees of success in reducing flood runoff rates depending (in addition to the magnitude and intensity of the rainfall and antecedent moisture conditions) on the percentage of the basin treated and basin topography. Generally, as drainage areas increase, upland treatment measures, wetlands, and small detention structures have had less effect in decreasing peak flowerets.

Upland wetlands restoration and upland land treatment practices can be effective for smaller floods but diminished in value as storage capacity is exceeded in larger floods such as the Flood of 1993. Present evaluations of the effect that wetland restoration would have on peak flows for large floods on main rivers and tributaries are inconclusive. For large events like the 1993 flood, upland treatments had little effect. Land treatment and detention storage (upland wetlands) can play a role in reducing peak runoff in some watersheds but are not a panacea for solving flood problems. The Review Committee concluded that only a combination of upland and floodplain management practices can reduce floodplain damages. Federal Insurance Programs.

At the time of the 1993 flood, the federal government operated two insurance programs that provided claims payments to those impacted by the Midwest flood; the National Flood Insurance Program (NFIP) and the Federal Crop Insurance Program. Table 1 summarizes claims payments from these programs by state. Under both programs, individuals paid an annual insurance premium to the government and the government provided insurance coverage.

Table 1 Summary of Federal Insurance Claims Payments by State for the 1993 Midwest Floods in Millions of Dollars.

Flood insurance coverage on buildings and their contents is available through the NFIP in participating communities. Under the NFIP insurance premiums for buildings that pre-date the identification of the flood hazard in a particular community are subsidized, but for buildings built after that date, premiums are based on full actuarial rates. All costs of administering the program, including the costs of floodplain mapping and salaries of federal employees are charged to policy holders. The Midwest flood was the third most costly in terms of NFIP payments, exceeded only by Hurricane Hugo and the December 1992 coastal storm that struck New York, New Jersey, Massachusetts, Delaware, and Connecticut. In 1993, over half of the losses and two thirds of the payments were in Missouri. States in the upper basin had lower average payments since buildings were generally subject to shallow flooding along tributaries which flooded basements and some first floors. States in the lower basin had much higher average losses reflecting the deep flooding in the bottoms along the main stems of the Mississippi and Missouri rivers). High average payments in Missouri also reflect large payments to small businesses and other non-residential buildings. Even in the counties with disaster status, in excess of 80,000 insured properties did not sustain flood losses. Some of these were behind levees that did not overtop or fail, but most were on tributaries that did not flood or where flooding was of less than 100-year frequency. At the time of the 1993 flood farmers could have protected themselves from actual crop losses or prevented planting caused by uncontrollable natural events through purchase of crop insurance from the FCIC. This government corporation within the US Department of Agriculture (USDA) provided coverage for 51 crops in the event of loss from drought, excess soil moisture, flood, frost, hail, wind, insects, and other natural perils. Historically drought has been the major cause of crop loss (55 percent) while floods represent only two percent of claims. Excess soil moisture, however, represents 16 percent of losses. Farmers must purchase the insurance early in the crop year. For example, a policy to cover a com crop planted in 1993 in the Midwest would have to be purchased by April 15. Farmers could have chosen the level of insurance coverage that they wished to purchase, but they were not able to insure their crop for the full value. Maximum coverage was 75 percent of expected crop yield. To encourage participation, the federal government subsidized crop insurance premiums up to 30 percent and paid administrative, actuarial, underwriting, and selling expenses.

The NFIP was created by Congress in 1968 in response to mounting flood losses and escalating costs to the general taxpayer for disaster relief in the belief that flood insurance is preferable to disaster assistance. To encourage participation in the NFIP by communities and purchase of flood insurance by individuals, the federal government subsidizes the premiums for buildings constructed prior to the issuance of a FEMA Flood Insurance Rate Map (FIRM). This subsidy also recognizes that many floodplain buildings were built or purchased without knowledge of the flood risk. New construction (post-FIRM) is charged an actuarial premium that reflects the property's risk of flooding. Approximately 60 percent of NFIP policyholders pay a full actuarial rate and 40 percent are subsidized. For the NFIP to be successful in indemnifying property owners from flood losses and reducing federal expenditures for disaster assistance, a high percentage of property owners must purchase and maintain flood insurance coverage. The program depends on the mandatory flood insurance purchase requirement contained in the Flood Disaster Protection Act of 1973 and voluntary purchase by other property owners at risk. The 1973 Act requires the purchase of flood insurance by property owners who receive federal grants or loans or loans from a federally supervised, regulated, or insured lender for the acquisition, construction, or improvement of structures located in identified special flood hazard areas (the 100-year floodplain). In the 9-state region affected by the 1993 flood, only about 20 percent of structures in the floodplain carried flood insurance, a rate well below optimal levels.

The NFIP has not achieved the public participation needed to reach its objectives. Estimates of those covered by flood insurance nationwide range from 20 to 30 percent of the insurable buildings in identified flood hazard areas. Estimates in the Midwest flood area ranged from below 10 percent up to 20 percent. None of the estimates are authoritative, since no nationwide inventory of floodprone structures exists. The Review Committee obtained reliable structure counts for a number of Midwest communities. Market penetration in these communities ranged from less than 5 percent to more than 50 percent. The Review Committee believed that market penetration in small rural communities was probably less than 10 percent. For most medium to large communities, market penetration appeared to be in the 20 to 30 percent range. For a few large communities with middle-income floodplain populations and a high degree of flood hazard awareness among community officials, lenders, and property owners, market penetration exceeded 30 percent and, in one instance, 50 percent.

A perception persists that disaster assistance compensates homeowners as fully as flood insurance coverage. A particular concern expressed by communities and others after the Midwest flood was that disaster victims, particularly those with lower incomes, who obtained disaster assistance from the Individual and Family Grant Program, the Disaster Housing Program, the Red Cross, and other programs ended up as well off as those who purchased flood insurance and received payment for claims. Generous disaster assistance creates negative incentives for the purchase of flood insurance.

The Review Committee recommended that the Administration take a number of actions to improve both the NFIP and the Crop Insurance Program. Many of these recommendations were incorporated in a Flood Insurance Reform Act signed by the President in late 1994. Several changes in the Crop Insurance Program have also been made by the Congress and the President, although, in both cases there still remains much room for improvement.

3.4 Relocations and Buyouts Following Floods

Throughout history, well-designed and well-sited structural measures have demonstrated their effectiveness in protecting property and saving lives. While structural measures have been the primary approach to flood damage reduction, another approach to minimizing vulnerability, not widely used in the past, is the removal of vulnerable populations from the floodplain. Because of the severity and duration of the 1993 flood, the general public took a new interest in this strategy. Building on its experience with the NFIP, the FEMA capitalized on this interest in removals. The Administration responded by targeting federal recovery that support buyouts and relocation of floodplain populations. The fundamental value of buyouts over structural approaches is that they completely eliminate flood risk for affected individuals and, at the same time, may have environmental and hydrologic benefits. Relocation associated with buyouts can, however, involve social, environmental, or hydrological impacts. The Administration established buyouts of flood-damaged properties as the first priority for mitigation funds available for the Midwest flood and, by October 1995, had approved over 8200 homes for voluntary relocation or buyout.

This approach represented a clear turning point in federal flood recovery policy, since it is the first time that buyouts have been attempted on such a large scale. Buyouts were an appropriate federal response for the Midwest flood and for floods like it. Many of the buyout neighborhoods had been damaged repetitively by flooding. Subject to deep and long duration flooding, they were isolated by floodwaters for extended periods of time. In addition a significant percentage contained older, lower value housing, much of it of poor quality and in need of rehabilitation. Under the right circumstances, buyouts will not only reduce flood damages and protect people and property but also achieve other objectives such as improving the quality of affordable housing, increasing recreational opportunities and wildlife values, and general betterment of the community.

During the 1993 flood, acquisition by federal and state governments of environmental easements or title to agricultural lands subject to frequent flooding also became tools in assisting recovery and in removing people from long-term flood vulnerability. In addition to meeting the needs of disaster relief victims, these programs were effective in achieving the nation's environmental goals. Since the 1993 flood, federal and state governments have obtained interest from willing sellers in over 100,000 acres of marginal bottomlands, and were additional funds available, would have obtained interest in over 60,000 more acres.

4. Environmental Implications of Human Intervention

Fish and wildlife resources in the upper Mississippi River Basin have been significantly affected by the loss of wetlands and other terrestrial and aquatic habitats due to construction for navigation and flood damage reduction structures.

The upper Mississippi River was originally a free-flowing, alluvial riverine environment with associated riparian habitats. Construction of navigation control structures (rock dikes) and installation of the slackwater navigation dams have created habitat types substantially different from those found in a free-flowing alluvial river. Habitat types within the upper Mississippi River slackwater navigation pools are created by coincident physical, water quality, and botanical characteristics. River position, depth, water-surface area, stage and discharge, vegetation, river-bottom types, water quality, and the superimposed structural elements within the river define the various habitats. Three distinct habitat zones occur in the slackwater navigation pools. The upper end of each pool is like the original river although subject to exaggerated water level fluctuations from the upstream dam releases. Marsh development is limited. In the middle portion of the pools, downstream impoundment backs water up and over the islands and old hay meadows, creating large areas of shallow water. This section has the best marsh development, and some deep sloughs and wooded islands can be found. In the lower end, immediately above each dam, wide open water lake-like areas occur. While impoundment of the upper Mississippi River for slackwater navigation created a variety of backwater and side-channel habitats, these dams also slowed river currents, starting the irreversible process of sedimentation. Many backwater habitats are filling with sediments from the erosion of upland agricultural and developed lands. Rock dikes and channel maintenance dredging also contribute to the problem. Mississippi River backwaters still provide critical fish production and nursery habitats, but may be lost to sedimentation and eutrophication within 50 years. Downstream from its confluence with the Missouri River, the upper Mississippi River takes on a very different character, similar to that of the Missouri River. Forty-six species of Mississippi River fish, virtually all of which have been affected by flood damage reduction measures and navigation, are listed by basin states as rare, threatened, endangered, or a species of special concern.

Parts of the Missouri River were well known as a braided river with swift, muddy flows. The historic floodplain was a ribbon of islands, chutes, oxbow lakes, backwaters, marshes, grasslands, and forests. Sandbars and wooded islands dotted the channel. Between 1879 and 1954, human actions and natural changes shortened the river by 45.6 miles, reduced river surface area by over 50,000 acres, reduced the number of islands from 161 (24,419 acres) to 18 (419 acres), and converted nearly 67,000 acres of river habitat from public to private ownership, most to agriculture. Nearly one-third of the Missouri River has been impounded, another one-third channelized, and the hydrologic cycle, including temporal flow volume and sediment transport, has been altered on the remainder. The Missouri River formerly had peak run-off during two periods, March-April and June. Prior to 1954 flushing flows, known as dominant discharge, occurred every 1.5 years. The river was in a state of equilibrium; net sediment entering a reach replaced an equal amount leaving allowing for ample habitat development, and aquatic nutrition. Loss of sediment load led to channel degradation which contributed to the loss of off-channel habitat and further severed the river from its floodplain. Since the early 1950s the Missouri River has thus been deprived of a floodplain in most reaches. Water temperature, photoperiod, and run-off cues have been altered by reservoir releases for navigation and other purposes are rare.

Aquatic and terrestrial habitats of the Illinois River Valley have suffered a series of cataclysmic events since 1900: (1) permanent rise in water level from water diverted from Lake Michigan, (2) the draining of more than half of the 400,000 acre floodplain through the construction of levees and pumping stations, (3) an upsurge in untreated urban and industrial pollution during the 1920s, (4) the creation of a 9 ft. channel and its attendant navigation dams in the 1930s, and (5) an acceleration in sedimentation rates following World War II, apparently resulting from an increase in the amount of open row crops grown within the basin. As an example, in 1908, a 200-mile reach of the Illinois River produced 10% of the total US catch of freshwater fish (employing 2,000 commercial fishermen and yielding 24 million lbs. of fish annually). Commercial fish yield totaled about 178 lbs/ac of permanent water, but by the 1950s yield had dropped to 38 lbs/ac and by the 1970's to 4 lbs/ac, totaling 0.32% of the total US freshwater harvest.

Alteration of Mississippi, Illinois and Missouri Rivers and floodplains has resulted in significant changes or losses of habitat. The disruption of natural ecosystems has caused the destruction of many native species populations and has caused an increasing number to be listed as threatened or endangered.

5. Management Aspects of Floodplain Management

5.1 General

Since passage of the Flood Control Act of 1936, the federal government has dominated the nation's flood damage reduction efforts and, as a result, the nation's floodplain management activity. Structural programs were deemed important and were also the principal sources of funds for any efforts to stem the rising tide of flood losses. In recent years, the federal government has begun to support nonstructural approaches. Many states, tribes, and local governments developed and carried out floodplain management efforts that both reduced flood damages and enhanced the natural functions of floodplains. There are more than 8,000 miles of levees in the Upper Mississippi Basin. Approximately half were constructed by the federal government or were locally constructed using federal standards. The levee system in the Upper Mississippi Basin is a loose amalgam of federal store and locally constructed levees. In carrying out these programs, however, they were hampered by uncoordinated and conflicting federal programs, policies, regulations and guidelines that have hindered efficient floodplain management. Some state and local governments have not been as active in floodplain management. With the federal government assuming the dominant role and funding most ecosystem restoration, flood damage reduction, and flood recovery activities, the incentive has been limited for many state, tribal and local governments, businesses, and private citizens to share responsibility for making wise decisions concerning floodplain activity.

5.2 Management Responsibilities

The division of responsibilities for floodplain management activities among and between federal, state, tribal, and local governments is not clearly defined. Within the federal system, water resources activities in general and floodplain management in particular lacks coordination. Management of the nation's water resources is provided by several federal agencies. Yet water resource issues are inextricably linked and accomplishment of agency mandates requires coordination and collaboration among agencies. The National Flood Insurance Act of 1968 required reports to Congress analyzing the implementation of current programs and recommending actions needed to achieve a unified program of planning and action at all levels of government to reduce flood losses and losses of floodplain natural values. Despite these Unified National Program for Floodplain Management reports, the United States, in practice, has no unified national program for floodplain management.

State and local governments have little fiscal stake in floodplain management; without this stake, few incentives exist for them to be fully involved in floodplain management. State governments must assist local governments in dealing with federal programs but, in may cases, do not become involved in federal-local activities.

In 1977 with issuance of Executive Order (EO) 11988, Floodplain Management, President Carter raised federal agency attention to issues of floodplain use. It was apparent following the 1993 flood that some federal agencies either were unaware of or misunderstood the requirements of the EO and either built or supported building in floodplains. Under the EO, federal agencies must demonstrate that no practicable alternative site exists outside of the floodplain, and if no alternative exists, take steps to minimize direct and indirect impacts of the proposed action and to restore and preserve the floodplain. Discussions with the FEMA, USACE, and state floodplain managers revealed several examples of apparent non-compliance by federal agencies with the EO. While the responsible agencies no doubt believed they had complied with the EO, these developments point out some of the deficiencies with the EO. Among the most notable examples were a low-income housing project funded by HUD and a federally funded state prison within floodplains, and a proposed construction of a federal weather station behind an uncertified levee.

Federal and state oversight over non-federally constructed levees was and remains diffuse. Several states regulated construction in floodplains, but many did not. The situation was further exacerbated by the potential for future flow increases that could occur if development continued upstream and by the uncertainty about changes that may occur in long term weather patterns. Few states controlled either the decision about where levees are placed relative to the river channel or whether a particular levee should be protected from overtopping (floodfought) during a flood, although such actions can have hydraulic and environmental consequence elsewhere. Some states had little or no involvement in the processes associated with federal levee programs since federal agencies generally dealt directly with local levee districts. The Review Committee recommended that the President should propose enactment of a Floodplain Management Act to establish a national model for floodplain management, clearly delineate federal, state, tribal, and local responsibilities, provide for fiscal support for state and local floodplain management activities, and recognize states as the nation's principal floodplain manager. He should also issue a revised Executive Order clearly defining the responsibility of federal agencies to exercise sound judgment in floodplain activities. In late 1994, a committee of the US Senate proposed a Floodplain Management Act, However, as a result of the change of control of the Congress in November, 1994, the action was tabled. Several states, on their own, have substantially revised their floodplain management programs.

5.3 Planning

The Review Committee concluded that the principal federal water resources planning document, Principles and Guidelines, was outdated and did not reflect a balance among the economic, social, and environmental goals of the nation. This lack of balance was exacerbated by a present inability to quantify, in monetary terms, some environmental and social impacts. As a result, these impacts are frequently understated or omitted. Many critics of Principles and Guidelines saw it as biased against nonstructural approaches and the that the federal government supports more structural than non-structural projects. The Review Committee recommended that the President should immediately establish environmental quality and national economic development as co-equal objectives of planning conducted under the Principles and Guidelines. Principles and Guidelines should be revised to accommodate the new objectives and to ensure full consideration of nonstructural alternatives. He should also direct that federal agencies give full consideration in planning to all possible alternatives for vulnerability reduction, including permanent evacuation of floodprone areas, flood warning, floodproofing of structures remaining in the floodplain, creation of additional natural and artificial storage, and adequately sized and maintained levees and other structures.

5.4 Management of the Upper Mississippi Basin System

At the time of the 1993 flood there was no coordinated strategy for effective management of the water resources of the upper Mississippi River Basin. Such a strategy still does not exist. Responsibility for integrated navigation, flood damage reduction and ecosystem management has been and remains divided among several federal programs. This is in sharp contrast to the lower Mississippi River Basin where one activity, the Mississippi River Commission provides for a comprehensive approach to water resources development within the alluvial valley of the Mississippi. The current flood damage reduction system in the upper Mississippi River Basin represents a loose aggregation of federal, local, and individual levees and reservoirs. This aggregation does not ensure the desired reduction in the vulnerability of floodplain activities to damages. Many levees are poorly sited and will fail again in the future. Without change in current federal programs, some of these levees will remain eligible for post-disaster support and will be repaired again after the next flood. Current federal rules essentially require, the federal government to repair these levees each time they fail.

The Review Committee recommended that the President assign responsibility, in consultation with the Congress, to the Mississippi River Commission, for integrated management of flood damage reduction, ecosystem management, and navigation on the upper Mississippi River and tributaries and assign the Commission responsibility for development of a plan to provide long-term control and maintenance of sound federally built and federally supported levees along the main stems of the Mississippi and Missouri Rivers.

5.5 Use of Science and Technology

At the time of the 1993 flood, the nation was not using science and technology to full advantage in gathering and disseminating critical water resources management information. Opportunities existed to provide information needed to better plan the use of the floodplain and to operate during crisis conditions. The Review Committee recommended that the Administration should establish a federal information clearing house to provide federal agencies and state and local activities the information already gathered by the federal government during and following the 1993 flood and to build on the pioneering nature of this effort, and exploit science and technology to support monitoring, analysis, modeling, and the development of decision support, data acquisition, and geographic information systems for floodplain activities.

6. In Sum

The Mississippi River Flood of 1993 was a significant hydrometeorological event that produced widespread impacts across the Midwest of the United States. Study of the events surrounding the flood as well as the flood itself, have produced a significant amount of information concerning the use and management of lands subject to flooding. The recommendations of the Interagency Floodplain Management Review Committee provide a focus for actions that should be considered as the United States grapples with the problems of reducing the vulnerability of floodplain activities and those who inhabit the floodplain.

References

American Red Cross National Headquarters, "Mid-West Floods 1993 American Red Cross Disaster Relief Operations, Statistical and Cost Report," (ARC, 1993).

Ayres, B. Drummond, Jr., "What's Left from the Great Flood of '93," The New York Times, (August 10, 1993).

Cassidy, Dan, and Rickert Althaus, "The Flood of 1993: The Economic Aftermath," Choices (First Quarter 1994), pages 29-31.

Bhowmik, Nani G., (ed.), The 1993 Flood on the Mississippi River in Illinois, (Illinois State Water Survey Misc. Publication No. 151, 1993).

Goolsby, Donald A., William A. Battaglin, and E. Michael Thurman, Occurrence and Transport of Agricultural Chemicals in the Mississippi River Basin, July through August 1993, US Geological Survey, Circular 1120-C, (Washington, DC: USGS, 1993).

Harper, Edwin L., President and Chief Executive Officer, Association of American Railroads, testimony before the Subcommittee on Transportation and Hazardous Materials, Committee on Energy and Commerce, US House of Representatives, September 23, 1993.

Hey, D., "Prairie Potholes," pages 505-509 in National Research Council, Commission on Geosciences, Environment, and Resources, Water, Science and Technology Board, Committee on Restoration of Aquatic Ecosystems: Science, Technology, and Public Policy, Restoration of Aquatic Ecosystems, (Washington, DC: National Academy Press, 1992).

Hillaker, Harry (1993). Special Summary, Great Iowa Floods, 1993. Iowa Department of Agriculture, Des Moines.

Klemme, Dale, Relocation Coordinator, Prairie du Chein, WI, personal communication, 1994.

Knight, D. Karen, US Environmental Protection Agency, Headquarters Emergency Operation Center, personal communication, August 25, 1993.

Kulik, Bernard, Associate Administrator for Disaster Assistance, Small Business Administration, personal communication, May 3, 1994.

Leopold, L., "Flood Hydrology and the Floodplain," pages 11-14 in Universities Council on Water Resources Update, (Carbondale, IL: UCOWR, Spring 1994).

Oliver, Sheila, Department of Interior, Bureau of Indian Affairs, Sisseton Agency, South Dakota, personal communication, May 16, 1994.

Satterlund, D.R., and P.W. Adams, Wildland Watershed Management, (New York, NY: John Wiley & Sons, Inc., 1992).

Shepard, Bonnie, Federal Emergency Management Agency, Federal Insurance Administration, personal communication April 14, 1994.

Smith, V. Kerry, "Nonmarket Valuation of Environmental Resources: An Interpretive Appraisal," Land Economics, 69(1): 1-26 (February 1993).

Ribando, Marc O., and Daniel Hellerstein, Estimating Water Quality Benefits: Theoretical and Methodological Issues, Technical Bulletin 1808, (Washington, DC: USDA Economic Research Service, September 1992).

Taff, Steven J., and Wilbur Maki, University of Minnesota Department of Agricultural and Applied Economics, (Letter Report, October 25, 1993).

Trilling, Donald R., Director, Office of Transportation Regulatory Affairs, US Department of Transportation, personal communication, February 16, 1994.

US Army, Corps of Engineers (1995), Floodplain Management Assessment, Washington, DC.

US Congress, PL 103-75, Emergency supplemental appropriations for relief from the major, widespread flooding in the Midwest for the fiscal year ending September 30, 1993, and for other purposes, (Washington, DC, August 11, 1993).

US Department of Agriculture, Soil Conservation Service data sheet, (Columbia, MO: USDA, October 1993).

US Department of Commerce, National Oceanic and Atmospheric Administration, National Weather Service, Natural Disaster Survey Report: The Great Flood of 1993, (Silver Spring, MD: DOC, NWS, February 1994).

US Executive Office of the President (1994), Sharing the Challenge: Floodplain Management into the 21st Century, Washington, D.C.

US Interagency Advisory Committee on Water Data, Hydrology Subcommittee, Guidelines for Determining Flood Flow Frequency, Bulletin #17B (Reston, VA: USGS, March 1982).

US National Performance Review. Department of the Interior: Accompanying Report of the National Performance Review, (Washington, DC: US Government Printing Office, 1993).

US Department of Transportation, Maritime Administration (MARAD), data sheet, August 17, 1993.

US Environmental Protection Agency, The 1993 Midwest Flood Water Quality Monitoring Update of EPA's Flood Monitoring Task Force/Water Workgroup, (Issue 2, December 31, 1993).

US Federal Emergency Management Agency, Federal Insurance Administration, computer print-out (Washington, DC: March 16, 1994).

US Federal Emergency Management Agency, "Big Disasters 1989-1994 - Projected Infrastructure Funding (formerly Public Assistance)," (Washington, DC: FEMA, April 1, 1994).

US Federal Interagency Floodplain Management Task Force, Floodplain Management in the United States: An Assessment Report, Volume 2, (Washington, DC: FIFMTF, 1992), pages 9-8.

Wilson, John, Tribal Liaison, Environmental Protection Agency, Region VII, personal communication, May 16, 1994.

Young, Admiral Frank, Director, Office of Emergency Preparedness/National Disaster Medical System, Public Health Service, personal communication, January 10, 1994.

ENDNOTES

1. This paper is based on and in part extracted verbatim from the report of the Interagency Floodplain Management Review Committee, which, under charter from the Executive Office of the President of the United States., conducted a six month review of the 1993 flood event. The report. Sharing ¡he Challenge: Floodplain Management into the 21st Century, was submitted to the White House in July 1993. The Review Committee consisted of federal engineers and physical, social, and biological scientists who contributed technical and institutional knowledge in the fields of flood damage-reduction and river basin ecosystem management. Of the 31 -member Review Committee, 15 members were located in Washington, DC, and 16 formed the Scientific Assessment and Strategy Team (SAST), which operated from the Earth Resources Observation System (EROS) center at Sioux Falls, South Dakota. The SAST was chartered by the White House in November 1993 "to provide scientific advice and assistance to officials responsible for making decisions with respect to flood recovery in the upper Mississippi River Basin." It was incorporated into the Review Committee in January 1994 to serve as its research arm for scientific analysis. Other members of the committee were Dr. Margriet Caswell, Thomas Wheri, Department of Agriculture, Richard DiBuono, Arnold Robbins, Harry Shoudy, Department of the Army (Corps of Engineers) Robert Clevenstine, Jerry Rassmussen, Department of the Interior, Shannon Cunniff, Joseph Ferrante, Lewis Rosenbluth, Environmental Protection Agency, and Mary Jean Pajak, Michael Robinson, Federal Emergency Management Agency and their work is reflected in this paper.

2. The accuracy of the assigned flood recurrence interval remains in question. St. Louis has experienced floods similar to the 1993 event in 1900, 1909,1927 and 1973.

Anexo 4.4/Appendix 4.4 - Programa de mitigación de inundaciones de la cuenca del río Magdalena-Colombia

por David Humberto Ojeda Awad y Héctor Guzmán Alvarez

INTRODUCCION

MOTIVACIONES Y PROBLEMÁTICA

Colombia ha venido buscando desde hace mas de seis décadas un marco físico de trabajo en donde se articulen orgánicamente las actividades económicas con la protección de las infraestructuras sociales y la conservación del sistema natural.

Desde los años cincuenta se ha probado la administración de los recursos naturales y el ambiente a través de entidades del orden publico basada en el aérea geográfica de la cuenca Hidrográfica. Estos entes, sin cumplir su objetivo de administrar los recursos naturales como bienes públicos en beneficio de la ciudadanía, asumieron paralelamente la responsabilidad de la gestión ambiental, incluyendo actividades de planificación, ordenamiento, y manejo y por ende acciones de preservación, conservación y protección. Para tal fin se comprometieron componentes de investigación, transferencia y ajuste de tecnología y fomento de los recursos naturales renovables. Todo lo anterior llevó a un incumplimiento en cada uno de los anteriores componentes y por ende a un desgaste en la acción Estatal.

Dentro de este esquema, la problemática de las inundaciones del rio Magdalena, las medidas de control se vieron expresadas solo en medidas estructurales, localizadas puntualmente; descuidando la visión integral que debe primar en cualquier actividad que busque resolver problemas ambientales.

La multiplicidad de funciones, componentes y elementos, al igual que la falta de claridad conceptual en el manejo ambiental, hicieron fracasar este modelo de abocar la problemática ambiental al igual que lo sucedido en el manejo político administrativo del país.

La promulgación de la nueva Constitución Nacional en el año de 1991, puso en marcha un nuevo esquema de relación Estado ciudadanía basado en el proceso de la participación comunitaria expresado en su forma política como Descentralización Político Administrativa y otorgamiento de autonomía local para el manejo de sus propios asuntos en concordancia con las políticas nacionales.

METODOLOGIA Y ACTIVIDADES

Colombia frente a la problemática de las inundaciones del Rio Magdalena y sus afluentes ha venido, a través de un proceso de gestión directa de Estado, ejecutando acciones a través de diferentes instituciones (Ministerio de Obras Publicas, Ministerio de Agricultura, Corporaciones Autónomas Regionales - CVC, Empresas de Energía y Acueducto de Bogotá). Estas realizaciones han sido puntuales, aisladas y sin una estructuración integral y sistémica, basadas solo en obras estructurales sin participación de la comunidad.

La conceptualización de cuenca y el manejo integral de los recursos naturales, a pesar de verse reflejada en la legislación desde 1974 (Código Nacional de los Recursos Naturales), no ha acompañado este proceso, perdiéndose de esta manera la visión integral y orgánica propia de un enfoque sistémico que aporta la conceptualización de cuenca hidrográficas para afrontar situaciones ambientales.

Con el proyecto Colombo Holandés se introduce el elemento de predicción a partir de sistemas de alarma instalados en redes especiales.

Otras medidas ejecutadas por el Estado han sido la realización de estudios de producción de sedimentos, comportamiento hidrológico de la cuenca e hidráulico del cauce principal del río, Diagnósticos Socio Económico del Alto Magdalena, al igual que investigaciones en el manejo de los recursos naturales (aguas, suelos, Bosques), transferencia de tecnología y fomento de los recursos naturales a través del proyecto Cuenca Alto Magdalena ejecutado por el Instituto de Desarrollo de Recursos Naturales (INDERENA), con recursos de crédito del Banco Mundial.

LOGROS Y LECCIONES

El país para la cuenca del Magdalena dispones de una Red de Básica Hidrometeorológica, y de una Red de Alertas conformada por cincuenta y nueve estaciones automáticas, las cuales son el elemento básico de prevención y mitigaciones de los efectos de las periódicas inundaciones que afectan las zonas bajas de las cuencas aportantes y de la planicie inundable del río Magdalena.

Resultado primario de la información producida son los mapas de amenazas por inundaciones que se vienen elaborando conjuntamente con la Dirección Nacional de Prevención y Atención de Desastres, al igual que una cartografía temática a escala 1:500.000 sobre la caracterización de sus recursos naturales.

De igual manera, para su administración integral y por mandato Constitucional se creo la Corporación del Río Grande de la Magdalena y por Ley 99/93 la constitución de trece corporaciones autónomas regionales (espacialmente cubren departamentos), encargadas además de la administración de los recursos naturales de la gestión ambiental y del acompañamiento del desarrollo sostenible de los trece departamentos que la componen, conforme a la nueva Ley de Plan de Desarrollo Nacional.

Finalmente, a partir de 1993 se creó el Ministerio del Medio Ambiente y el Instituto de Hidrología, Meteorología y Estudios Ambientales (IDEAM), encargados de aportar elementos del conocimiento ambiental hacia el logro de un ordenamiento ambiental territorial de la cuenca y orientar el manejo de los recursos naturales a fin de obtener el desarrollo sostenible, paradigma constitucional y marco de referencia del actual Plan de Desarrollo gubernamental EL SALTO SOCIAL.

1. La Cuenca del Río Magdalena

El río Magdalena es el mas importante de Colombia, situado en la parte céntrica del país atravesando de norte a sur la región Andina. Nace en el Páramo de las Papas en el Macizo Colombiano, en la laguna de la Magdalena localizada en el departamento del Huila y recorre al país de norte a sur en una extensión de 1.536 km, con un promedio de caudal de 7.100 m³/s.

El área tributaria de la cuenca es de 257.000 km² que corresponde al 22.8% de la superficie total del territorio nacional. Su recorrido se inicia dentro un valle interandino ubicado entre las cordilleras Central y Oriental y su afluente mayor, el río Cauca, entre las cordilleras Occidental y Central. El río atraviesa el corazón económico del país, en donde se encuentran las regiones más ricas, abundante en recursos naturales y las mejores condiciones de desarrollo, como son las zonas más industrializadas, mejor infraestructura vial, las más importantes desarrollos hidroeléctricos y áreas de desenvolvimiento agrícola importante. (Figura No. 1). La Cuenca Magdalena-Cauca está situada en la zona ecuatorial la cual determina su gran riqueza hídrica en donde se presenta una gran actividad climática en cuanto a precipitación con una media anual de 2.000 mm. con una variación dentro de la cuenca entre 800 mm. y los 5.000 mm. en algunas zonas. Considerando lo anterior, el caudal específico o rendimiento anual es de 27.3 L/S/km². (Ver cuadro 1).

El río Magdalena desde su nacimiento en la Laguna de La Magdalena hasta antes del embalse de Betania (hidroeléctrica situada aguas arriba de la ciudad de Neiva), presenta en su primer recorrido un régimen torrencial por su fuerte pendiente, la cual va disminuyendo mostrando características de un cauce meándrico y encajonado en donde se observa material de acarreo compuesto de piedra, grava y arenas gruesas. Esta misma situación de conserva hasta la ciudad de Honda en donde el río inicia una transición de un río meándrico a un río trenzado, es decir, el río amplía su lecho mayor para correr más libremente formando un sinnúmero de islas y ciénagas. Este tramo se le define como el Magdalena Medio. En la ciudad de El Banco el río sufre una nueva transición al penetrar una zona de características especiales la cual se le denomina la Depresión Cenagosa o zona inundable por ser una vasta llanura que se caracteriza por los numerosos caños y ciénagas que se entrelazan formando una red de canales a lo largo ancho de este sistema. Esta importante área sufre de inundaciones periódicas mayores durante los meses de octubre y noviembre, afectando cerca de 2 millones de hectáreas en los eventos más severos y unas 730.000 hectáreas en los eventos menos severos. (Ver Cuadro 1 y Figura No. 1).

Cuadro 1

De treinta y cuatro entes administrativos regionales conque cuenta el país para la administración y manejo de los recursos naturales renovables y el ambiente, trece lo hacen en la Cuenca del Magdalena, mas un ente de ámbito constitucional, la Corporación Autónoma del Río Grande de la Magdalena, encargada de la recuperación de la navegación, de la actividad portuaria, la adecuación y conservación de tierras, la generación y distribución de energía y el aprovechamiento y preservación del ambiente, los recursos ictiológicos y demás recursos naturales renovables.

2. Problemática de las Inundaciones

La cuenca del río Magdalena presenta toda una gama de condiciones hidrológicas y todas las clases de ríos conocida; esta situación se debe a su posición geográfica y su orografía en donde se localizan tres grandes cordilleras con sus altiplanos y grandes planicies aluviales. La red hidrográfica de la cuenca es abundante en donde se presentan todos los afluentes del río Magdalena con sus sistemas de ríos de origen torrencial y ríos de planicie que hacen que se presenten situaciones de crecientes o avenidas que van a producir inundaciones en sus zonas aledañas.

En la cuenca se puede establecer dos tipos de inundaciones de acuerdo al régimen de los cauces: inundaciones de tipo torrencial e inundaciones de tipo aluvial (de planicie).

Las inundaciones de tipo torrencial se originan en el área montañosa de toda la zona Andina correspondiente a la cuenca Magdalena-Cauca en donde existe una conformación topográfica abrupta, con fuertes pendientes, que durante los inviernos las grande precipitaciones transforman en peligrosos los diversos cauces, lo que conlleva a una amenaza latente de crecidas que ocasionan desbordamientos en sus márgenes y la misma velocidad del agua puede producir grandes daños en zonas pobladas.

Otro fenómeno que está relacionado con las crecientes es el de las avalanchas, que son ocasionadas por deshielo o represamiento ocasionado por el taponamiento del cauce debido al deslizamientos de laderas en donde los suelos son inestables y no muy consolidados. El taponamiento provoca el embalsamiento de grandes volúmenes de agua y cuando la presa natural falla, estos volúmenes fluyen con gran fuerza aguas abajo produciendo erosión en las laderas con arrastre de piedras y sedimentos, causando por consiguiente grandes estragos y destrucción.

Estos fenómenos de origen torrencial son los más frecuentes en la cuenca durante las épocas de lluvias y los que producen las mayores pérdidas materiales y de vidas humanas.

Las inundaciones de tipo de planicie son las que ocurren en las áreas bajas de los afluentes y en las zonas inundables incluyendo la depresión cenagosa del río Magdalena. Los valles del río Magdalena y de su afluente el río Cauca están sometidos a inundaciones de larga duración, como se mencionó en el aparte anterior; en donde las crecidas son más reguladas y no tienen una incidencia en la propia magnitud del caudal y solamente tienen importancia a las que obedecen únicamente a lluvias prolongadas dentro de la cuenca y obedecen a los períodos de invierno. Estas crecidas así producidas son inicialmente lentas y cuando exceden la capacidad del cauce y de las propios vasos amortiguadores como las ciénagas, se producen los desbordamientos inundando las zonas ribereñas.

El régimen de lluvias en la cuenca del río Magdalena es básicamente bimodal con eventos máximos en abril, mayo y octubre, noviembre. Los mayores caudales se presentan en el segundo período del año en donde se reduce considerablemente la capacidad de almacenamiento. El almacenamiento total de las zonas bajas con ciénagas se ha estimado en 40.000 millones de m².

En Colombia, como consecuencia de avenidas, inundaciones y otros eventos hidrometeoro-lógicos extremos, se registran anualmente perdidas materiales por valor en promedio de 35 millones de dólares, más de 65000 personas damnificadas y alrededor de unas 150 de vidas humanas perdidas. Las estadísticas indican que las pérdidas podrán alcanzar unos 70 millones de dólares y 230 víctimas en promedio una vez cada 20 años. La cuenca Magdalena-Cauca son las área más vulnerables del país, contando el 90% de los daños y el 70% de pérdida de vidas humanas. Además, las actividades agrícolas, la industria manufacturera y el aprovechamiento de los recursos hídricos, padecen grandes pérdidas materiales indirectas.

En el año de 1994, las inundaciones produjo pérdidas cercanas las 8.000 viviendas y más de 250.000 personas damnificadas que corresponde al 47.5% y el 80.7% respectivamente, del total producido por otros eventos de carácter catastrófico. Esto demuestra que las inundaciones son el evento más importante dentro la afectación a la economía del país. Ver cuadro 3 y figura 5).

Cuadro 3 Pérdidas por inundaciones en Colombia

De acuerdo a estos interrogantes se desarrollaron varias actividades entre ellas las más importantes que son el estudio y análisis hidrológico y el estudio hidráulico sobre el comportamiento dinámico del río mediante la concepción, montaje y calibración de modelos matemáticos que permitieran conocer la evolución de la incidencia de obras de protección contra las inundaciones sobre el régimen natural del río.

Los resultados obtenidos establecieron que las mejores soluciones estaban en aumentar la capacidad de amortiguación de las crecidas mediante la construcción de embalses de planicie o campos de inundación controlados.

El comportamiento natural del río dentro del sistema de amortiguación por medio de ciénagas en la zona lacustre del río Magdalena, el llenado y vaciado de estos vasos naturales ocurre en una forma sincronizada cuando se presentan los estados de crecientes y de aguas bajas. Esto indica que la ciénagas normalmente se están llenando antes que el río alcance su capacidad de descarga máxima sin desbordamiento. Por otra parte la evacuación de las ciénagas es lenta por la falta de capacidad de los caños de descarga.

Ahora, para aumentar la capacidad de amortiguación de las ciénagas se puede hacer controlando artificialmente las entradas y salidas de las mismas y aumentando su capacidad de almacenamiento por medio del confinamiento con diques. Esta solución por medio de estos campos de inundación controlados combinándolos con diques marginales en algunos tramos del río pueden disminuir considerablemente los desbordamientos en zonas de interés económico. Para el río Magdalena sena la solución más económica para mitigar el impacto de las inundaciones en toda el área afectada.

Para la protección de áreas urbanas localizadas en zona de inundación, se recomendó el confinamiento perimetral por medios de estructuras como diques y muros de contención y obras complementarias de evacuación de aguas usadas y lluvias. Sobre este caso se llevó a cabo un ejemplo de protección en la ciudad de Magangué localizada en la depresión cenagosa sobre el cual se contemplará en detalle más adelante.

Como se puede ver las soluciones anteriores se pueden catalogar como medidas estructurales, sin embargo el estudio recomienda medidas no estructurales como son los sistemas de predicción de crecientes por medio de avisos oportunos (Sistemas de alarma). Este tema en especial se tratará en capítulo aparte por ser la medida mas desarrollada hasta el momento en el río Magdalena.

Protección de la ciudad de Magangué. La ciudad está situada en el departamento de Bolívar, propiamente en la zona baja inundable del río Magdalena, sobre la margen izquierda del Brazo de Loba y una distancia de 230 kilómetros aguas arriba de Barranquilla.

Gran parte de la zona urbana de la ciudad se encuentra localizada sobre la zona aluvial del río la cual estaba sujeta a frecuentes inundaciones. En octubre de 1975 se presentaron una de las mayores inundaciones de los últimos tiempos presentándose niveles que alcanzaron a cubrir gran parte de las calles con alturas hasta de un metro por encima del suelo.

Con relación a lo anterior se realizaron los estudios correspondientes de protección de la ciudad según las recomendaciones del Proyecto Colombo-Holandés. Y luego durante los años de 1978 y 1979 se elaboraron los diseños y la construcción de las obras previstas.

Antes de la construcción de la protección, la ciudad de Magangué, el río Magdalena desbordaba frecuentemente causando inundaciones afectando gran parte de la ciudad perjudicando el centro comercial y un área densamente poblada y también el desbordamiento de ciénagas aledañas con el consiguiente anegamiento de las zona más pobres de la ciudad.

Se construyó un dique de tierra perimetral en una longitud de 6.796 metros y el realce del muro frontal existente componente del malecón - muelle de la ciudad. Además se construyeron cerca de 1.000 metros más de muro de contección.

Para solucionar la evacuación de las aguas negras y lluvias, se construyeron dos plantas de bombeo que son operadas por el municipio. Se adecuaron algunas ciénagas para el tratamiento de las aguas usadas para evitar la contaminación de las aguas del río Magdalena.

Esta experiencia hasta el momento ha dado resultados satisfactorios por cuanto la ciudad adquirió una nueva vida sin la preocupación de que sea afectada nuevamente por el problema de las inundaciones.

3.1 Estructura Organizacional

El desarrollo socioeconómico del país y la necesidad de reducir cada día el número de víctimas y pérdidas que dejan las temporadas invernales ha hecho forzoso que cada vez sea necesario realizar un trabajo más exhaustivo, mediante el cual se tengan en cuenta otras áreas del territorio nacional y también otros aspectos diferentes de las inundaciones. En la actualidad, el Instituto de Hidrología, Meteorología y Estudios Ambientales IDEAM pretende enfocar sus análisis a sistemas preventivos en aspectos tales como los incendios forestales, el fenómeno de las heladas, sequías, tormentas de marea y advertencias en temporada de huracanes, entre otros fenómenos de tipo natural ambiental.

En los comienzos de la era de la Alertas en Colombia, los técnicos y profesionales del antiguo HIMAT, una vez detectaban un fenómeno extraordinario que pudiese afectar a un sector del territorio nacional dentro de la Cuenca Magdalena-Cauca emitían un comunicado a los principales cuerpos de Socorro y a los medios de comunicación y mantenían coordinación y comunicación permanente sobre la evolución del fenómeno.

A raíz del desastre del Nevado del Ruiz, que conmovió al mundo entero en el año de 1985, el país entró en una etapa de concientización y reorganización de los sistemas de prevención, creándose el Sistema Nacional, dentro del cual posteriormente se creó la Oficina Nacional para la Atención y Prevención de Desastres, hoy Dirección Nacional para la Prevención y Atención de Desastres de Colombia, cuya tarea ha sido la de coordinar las actividades que realizan las diferentes entidades del orden nacional, departamental y municipal, no solo en materia de prevención sino de atención y durante la etapa de reconstrucción. El IDEAM hace parte del Comité Técnico Nacional con que cuenta esta organización y asesora a la Dirección en diferentes campos de la etapa de prevención básicamente, una de ellas a través de la Red de Alertas Ambientales.

Es a través de la Dirección Nacional para la Atención y Prevención de Desastres que el IDEAM efectúa la emisión de Avisos y Advertencias, Dirección que a su vez asume la responsabilidad de coordinar las acciones pertinentes según la calidad de la emergencia.

Con base en los estudios técnicos realizados hasta el momento, los cuales incluyen la utilización de fotografías aéreas e imágenes de radar y de satélite así como levantamientos topográficos y trabajos de comprobación de campo y gracias a la recepción de datos horarios vía satélite, radio y teléfono ha sido posible obtener puntos válidos de referencia que nos sirven de orientación para poder sugerir la declaración de un aviso, advertencia o Alerta. Eventualmente se ha recurrido a la aplicación de técnicas de simulación matemática mediante modelos de computador, lo cuales continúan siendo, en nuestros días, objeto de análisis.

Una vez efectuado el procesamiento y análisis de los datos y cuando las condicione lo ameritan se produce un comunicado sugiriendo, según la magnitud del evento, la condición de Alerta Amarilla (Aviso), Alerta Naranja (Advertencia) o Alerta Roja(Punto crítico de desbordamiento).

Sistema Nacional para la Prevención y atención de Desastres de Colombia

3.2 Alertas Hidrometeorológicas en la Cuenca del río Magdalena-Cauce

Desde finales de la década de los 70 se dieron los primeros pasos para el establecimiento en Colombia de una Red de Alertas hidrometeorológicas, bajo la asesoría de una misión canadiense la cual después de análisis minuciosos, recomendó iniciar la observancia sobre el comportamiento de los principales parámetros que inciden en las variaciones del tiempo y de los niveles de algunas estaciones de los ríos a lo largo de la cuenca Magdalena-Cauca.

La Red de Alertas hidrometeorológicas se basó inicialmente, en la instalación de una serie de radios para transmisión vía HF con centros especiales de acopio, de acuerdo con las características topográficas de nuestro territorio, y se inició el apoyo de las actividades propias del programa, gracias a la colaboración de algunas instituciones del orden nacional, regional y local tanto públicas como privadas, las cuales suministraban y siguen suministrando información diaria por teléfono o fax. Las experiencias logradas a través del ejercicio operativo de la red, permitieron establecer dificultades de seguimiento de los fenómenos extraordinarios en el momento oportuno pero facilitaron el mejoramiento sobre el conocimiento de algunas técnicas de observación y comunicación.

· Red actual

La aplicación de nuevos métodos de análisis y el acceso a tecnologías mejoradas han permitido al país avanzar en exactitud y oportunidad de los sistemas de prevención, gracias a la obtención permanente de datos en tiempo real y de imágenes periódicas que nos informan sobre el comportamiento atmosférico e hidrológico.

En la actualidad el IDEAM cuenta con 65 estaciones automáticas que transmiten información hidrológica y meteorológica durante las 24 horas del día, dos de las cuales son sobre aspectos mareográficos en el Pacífico; 20 estaciones distribuidas en la Sabana de Bogotá y tres en el altiplano nariñense para detección de heladas, mediante las cuales se recibe información por teléfono y radio. Se cuenta además con 30 observatorios meteorológicos para mediciones en superficie, cinco observatorios para la medición de la alta atmósfera, enlace con un sistema mundial de telecomunicaciones meteorológicas y una estación terrena para recepción de imágenes satelitales meteorológicas en baja resolución.

En la cuenca Magdalena-Cauca se han instalado hasta la fecha 59 estaciones automáticas dotadas de plataformas para transmisión por satélite distribuidas así.

Las funciones generales del Subprograma de Análisis y Aviso es analizar la información recibida del Subprograma de Recepción de la Información, efectuar el análisis estadístico en tiempo y espacio, producir el pronóstico por medio de modelos hidrológicos, asumir acciones respecto a Alertas según lo demande la evolución de eventos extraordinarios, producir los comunicados respectivos y mantener contacto directo con la Dirección Nacional para la Atención y Prevención de Desastres.

Las funciones específicas son:

· Diseñar la expansión y ampliación de la red de estaciones de acuerdo a los requerimientos

· Definir políticas en materia de uso, intercambio y manejo de la información de la Red de Alertas.

· Actualizar y mantener al día el Archivo Maestro.

· Interpretar los informes de diagnóstico de la información Automática y contribuir a la programación de las actividades de campo, tendientes a identificar la ubicación de plataformas DCP y las demás que demanden las circunstancias inherentes al programa.

· Impartir instrucciones al Subprograma de Recepción de Información sobre la observación en temporada especial o en situaciones inminentes de eventos extraordinarios de inundaciones.

· Efectuar el análisis en tiempo y espacio mediante el manejo de series estadísticas y otros de carácter físico.

· Elaborar Comunicados Especiales en tiempo de riesgo inminente e informar constantemente a la Dirección Nacional para la Atención y Desastres sobre el estado del río cuando está en observación especial.

· Ejecutar e interpretar los modelos de pronóstico hidrológico y mediante sus resultados en un evento extraordinario, emitir los informes correspondientes.

· Verificar la calibración de los modelos de pronóstico.

3.3 Medidas no estructurales de prevención

La ocurrencia de estos fenómenos de inundación, generalmente de gran extensión superficial, afecta no solo áreas urbanas sino también a grandes zonas agrícolas como también en áreas que son afectadas por avenidas de origen torrencial en los piedemontes de las cordilleras ocurridas en los afluentes directos del río Magdalena.

Para estos casos como medida de mitigación y prevención, el IDEAM adelanta un programa de mapas de amenazas en la cuenca Magdalena-Cauca con las entidades interesadas.

El primer proyecto se realizó en la cuenca del río Combeima, cuyo comportamiento de origen torrencial, presenta grandes avenidas causando inundaciones en zona urbana de la ciudad de Ibagué, localizada en el piedemonte de la cordillera Central zona alta del valle del río Magdalena.

Se elaboraron mapas de amenaza a escala 1:2.000 con períodos de retomo para 2, 20 y 100 años. Estos mapas es la base fundamental para llevar a cabo la zonificación de áreas de uso, para una planificación y reglamentación específica.

En la presa de Betania localizada en la parte alta del río Magdalena se elaboraron los términos de referencia para la determinación de mapa de amenazas aguas abajo de la presa.

Al construirse esta obra se creo un factor de riesgo para las zonas desarrolladas aguas bajo de la presa; el elemento principal y más frecuente es el paso de las crecientes por el embalse el cual constituye el riesgo que requiere el análisis relativo al comportamiento de estos caudales sobrantes aguas bajo dependiendo del manejo adecuado del embalse.

El propósito como resultado de lo anterior, es obtener los mapas de amenazas respectivos de las zonas aledañas al río que sean susceptibles a inundación, los cuales permitirán conocer los diferentes niveles de inundación y su frecuencia de ocurrencia asociada.

El mapa de amenazas planteará un esquema detallado para un trayecto de 160 kilómetros para todas las zonas de posible inundación con curvas de nivel cada 0.50 metros hasta la cota máxima que cubra un área para un período de 100 años. El mapa se elaborará a escalas 1:10.000 y 1:2.000 para zonas urbanas y se resaltarán las áreas de posible inundación para crecientes a filo de agua del embalse con períodos de retorno de 2, 10, 20, 50 y 100 años.

Como conclusión se establece que las medidas de protección más conocidas y de uso frecuente en años pasados son las estructurales pero muchas veces es mejor, sin embargo, que la mayor y mejor medida de protección y mitigación puede ser no solo el control de aguas de inundación sino el uso apropiado y administrativo de áreas susceptibles a inundación o a través de la zonificación y control de uso del suelo. A la larga este último esquema puede resultar menos costoso que los métodos de control de aguas.