This report was prepared under contract with the OAS by David Lang (Dominica).
‘Erosion’ is natural forces at work. The word erosion is an inclusive term for:
Erosion by running water is usually divided into several phases, some of which have names which are reasonably well known:
sheetwash describes erosion where water flows across land more or less in a sheet, and picks up particles of soil and carries them away. Its effects are often not very obvious to casual observation and it grades into
rilling is where flowing water makes very shallow sometimes broad channels across land and also picks up and wears away soil. Both sheetwash and rilling leave patches of deposited material wherever the gradient slackens and the slower flow will not carry all the load. When there is heavy rain and a large load is carried, deeper deposits form, often where the flow crosses a flat track or road and so becomes much more obvious.
gullying is where quite deep and relatively narrow channels are formed, by water running off the land. These may become very deep and large like the St. Kitts ghauts, and cut below the soil into underlying soft rocks. Gullies sometimes develop from the joining of rills but more often start because the existing natural drainage system has become overloaded as a result of a change of land use or a massive burn etc. Gullies usually start short, but may lengthen quickly with the head of the gully moving back up slope. At the lower end gullies may produce a deep fan of outwash material.
stream cutting: established streams have courses which change all the time. For long periods there may appear to be very little change but whenever the river flows really full—full to the top of its banks or overflowing—it will scour its bed and cut away at bends.
All of these are normal natural processes, but human activity often multiplies them greatly, both in frequency and size. Sheetwash and rilling are not common under natural vegetation in these islands, except where fire or hurricane has bared the land, but they have become very common on certain soils when farmed. Gullying is uncommon under natural vegetation but where vegetation has been disturbed by hurricane or burned off, which can happen naturally in volcanic territory and may explain the ghauts, gullying will occur.
Mass movement—landslide and rock fall—is relatively common under natural vegetation in the steeper landscapes of the Lesser Antilles. There are several different types of landslide, partly dependent on the kind of materials involved, but we can generalise by saying that most landslides (not rock falls) occur when, or just after, a mass of porous material is saturated or super-saturated with water. Prolonged heavy rain brings landslides. Rock falls may be triggered by rain in the odd case, but seismic events and gradual loosening under weathering are more common causes. However human activity once again multiplies the frequency and size of these events.
Wind erosion, where the wind picks up and carries light small soil particles (especially organic matter) and bounces or rolls larger particles along the land surface, cannot occur under a vegetation cover and is only found naturally in the formation of dunes in dry sandy beach or desert areas, or where some other factor has left a bare surface.
Human activity is usually effective in the following main ways:
The hazards caused by erosion include those caused by the deposition of eroded material, which can have a greater impact on land use and human activity or at least a more obvious impact than the loss of soil.
Rill and sheetwash erosion is mainly of consequence for farmers. Surface erosion mainly affects the part of the soil most valuable for farming, the organic topsoil. Losses of one or two centimetres—the maximum likely loss over a few years under normal conditions here—will probably not bother anyone else, but make a critical difference for farmers.
Deposition of fine materials (mud) at the foot of a slope on a farm may destroy crops and make tillage more problematic. By contrast a float of mud on a road may mean a few hours of inconvenience and some work for a machine—hardly a major problem.
Gully erosion could be more serious for other land users than for farmers, through the need for costly controls to avoid further expansion and the disruption of works already underway. Roads for example may be cut. Farmers may be able to contain gullies by using simple bush dam structures, filling and covering with vines etc., but for construction sites some kind of well-planned diversion may be required.
Mud and sand deposits at the outfalls of gullies, probably at the foot of a slope where the angle suddenly becomes more gentle, may be deeper than those from rills and gullies and more difficult to deal with.
Stream bank erosion need not be a great problem—it is simply natural adjustment to the course of the river—unless structures are built within the banks of the river or very close to them. It is a problem for some bridge footings, which must be built near the banks of rivers and streams. The effects of deposits from stream bank erosion should be insignificant, since they will either occur within the stream course, on natural levees or as shallow deposits on terraces, for which farmers at least are prepared. If a meandering river cuts a new course, on its flood plain, the area at hazard is somewhat more extensive. However it should be clear that such areas are subject to flood damage even apart from new stream courses. For any infrastructure construction in these areas, there needs special justification and planning.
Wind erosion is usually very damaging to young farm crops, because of the cutting action of particles whipped along by the wind. Tall, well-developed crops provide their own defence, to a great extent. Deposition is governed by wind direction and may take place anywhere. Probably in our conditions only a little material ‘flies’ in suspension and most ‘bouncing’ and rolling particles will be halted by the nearest bush or building. On the whole, the consequences are not likely to be very large.
Landslides and rock falls can be dangerous. Landslides have engulfed villages and killed people on many occasions in the Lesser Antilles, and there have also been many narrow escapes. Most landslides and rock falls contain relatively small quantities of material which moves only a short distance, but they can still present a threat to life. In steeper lands where there is potential for long distance travel, a long fall may provide the momentum to ride up an opposing slope. Flow slides, which are primarily found in allophane soils and little-weathered pumiceous ash, can travel long distances down quite gentle slopes. In areas where the more dangerous sorts of landslide / rockfall are possible (part of the areas of highest hazard class), it will be worthwhile to check whether buildings are in a potential landslide or rock fall path and monitor them. The hazard should promote caution but not be allowed to cause an excessive reaction. Many buildings in Dominica, a much more landslide-prone island than either Antigua or Barbuda, which appear to be prime candidates for landslide damage have existed with the potential threat for many years.
The best mitigation takes the form of avoiding, as far as possible, the practices which are conducive to erosion, and using defences only where that is not practical.
The maintenance of a good vegetation cover, preferably forest, is the best insurance against rainfall-based erosion and landslides. Dense grass cover, if tall, is nearly as good, unless it is tussocky like lemon grass, but it offers little protection from landslides. The most vulnerable conditions of the land are where extensive hillsides are completely cleared, something that occasionally happens when a major land use change takes place, but which can be very largely mitigated if standard conservation measures are used.
Attention needs to be paid to land use management. Farmers may or may not need to be instructed in conservation, but more investment in ensuring that good conservation practices are employed is necessary, to minimise both loss of fertility and the problems created by sedimentation.
Attention also needs to be given to Town and Country Planning. Concern expressed over reworking of stream banks, and natural excavation of stream courses by the stream in spate, indicate that planning has been neglected. Building roads across an occasional stream without providing adequate culverts and permitting house construction within a gully or on the edge of vertical stream banks in weak materials is not good practice. Mitigation of earthquake-induced rock falls and landslides also requires more careful recognition of potential recipient sites during Town and Country planning.
The main problem with dealing with water erosion is that land use is changed without appropriate conservation measures being taken. For example, converting a piece of land under bush to dwelling house lots results in a marked reduction in infiltration (the way water enters the soil) and a marked increase in runoff. The surface drainage of the area needs to be considered and linked to existing drainage in such a way that the drains will accommodate the flows produced by storms under the changed land use. This is a matter where planners need to enforce some rules. The proposed exit of St Kitts from the sugar industry will present an enormous problem unless land use change is well planned, as the soils presently protected by sugar cane have a high potential for erosion. The second problem relates to the design of drainage systems in general. Drainage systems need to be planned to deal with a whole sub-catchment. Those that do not, and do not integrate all factors, can cause more trouble than no system at all. Although no particular problems were evident during field visits to St. Kitts, except the case of the road to the southeast peninsula often mentioned previously, the alignment and construction of roads and particularly the provision for road drainage can provide long term instability problems and have been the initiator of major landslides and gullying throughout the Windwards. Again, this is a matter of planning, in which the planners need to ensure that developers including Public Works and farmers have given the matter adequate consideration.
There is no lack of knowledge within government departments of the means—planning, specific conservation measures—by which rainfall-induced erosion can be mitigated. But, for want of urgency, it is being neglected.
Wind erosion is only serious where literally bare soil is exposed to strong wind. The serious examples in Europe and North America ("dust-bowls") have occurred where large expanses of soil have been completely bare. In Europe, areas that had grown grains in fairly small fields without real problems were badly damaged by wind after the removal of hedges. Here any form of strip cultivation (provided it is not continuous over long distances, so as to provide a funnel), should be effective. This is a problem for agriculturists to tackle, but one which will affect others, so planners should be involved.
Mass movement under gravity is not something which can be easily stopped. Some land use practices increase the hazard of mass movement– removal of trees and loading weak slopes (with buildings or roads) for example– and should be proscribed as a matter of common sense, since we all have to pay for the resulting damage. Building on sites in the recipient area of such movements should be discouraged and would normally be a matter for rejection where Environmental Impact Assessments (EIA) are required. Again, planners should be able to minimise these problems.
Rock falls, simply a movement of loosened lumps of material under gravity, can only be mitigated by:
Landslides, (and some rock falls) may be mitigated by:
Stream courses can be canalised, as they often are where they pass through towns, but to prevent the adjustments which are natural processes of stream development will often be unreasonably expensive elsewhere and sometimes very difficult in most cases in St. Kitts, except perhaps downtown. It might be less difficult in Antigua but would still be costly and rarely giving a beneficial return commensurate with cost. The task is for planners to provide the right information.
These problems can really only be mitigated for the long term by:
Temporary protection can be effected by:
But it is unlikely that these will be cost effective except in a few instances.
Because of the tendency of most agriculture to accelerate natural erosion processes, erosion has been the subject of observation in relation to agriculture. Observations have been of two kinds:
These observations have led to the development of mathematical models used to estimate erosion losses or potential erosion. The main concepts are that each identified soil type has an intrinsic susceptibility to erosion (erodibility), that each rainstorm has a specific capacity to erode (erosivity) which can be assessed, and that land cover / vegetation has a major modifying effect on erosion potential.
As a land forming process, however, erosion is also very important in geomorphology, for its role in the development of slopes and stream courses and thus of natural landscapes.
The models used in this study have been developed using the general understanding that has developed of the processes involved and the role of various land characteristics in these processes. They are simple models using the following land characteristics as variables to estimate the effects of the apparent processes:
Models similar to these are widely used in the absence of local experiment and the kind of local data which engineers might provide.
The data used in these models is often a simpler version of what appears to be the appropriate data, which are used because they are available and the "better" data are not. Some better data can and should be obtained.
Erosion hazards are most commonly defined in relation to open land either under agricultural use or intended for an agricultural use. The application to development sites where some kind of construction is intended is related to localised incidents rather than a general picture. This emphasises the need for data at a more appropriate scale, which are not available at present.
The final map for St. Kitts combines gullying and mass movement (landslides and rock falls) and, where possible, an indicator of the probability of stream bank erosion. This is done by adding together the land unit by land unit values for gullying and mass movement and indicating separately those catchments in which stream bank erosion is highly likely.
The two final maps comprise a main map which is a composite and a supplementary , derived from the whole set of individual "erosion" process maps for each island. The full set, shown below and available as larger maps, comprises:
St. Kitts
Composite Maps: 1) Composite erosion hazard classes (gullying and landslide / rock fall hazard of bare soil areas except in the area of the Forest Reserve, where the hazard under vegetation is taken). 2) Stream bank erosion hazard , to be used as an annotation to map (Wind erosion and sheet and rill erosion are not considered of importance in building development in St Kitts)
| USAID/OAS Post-Georges Disaster Mitigation: http://www.oas.org/pgdm | Page last updated on 10 Sep 2001 |