The Arenal Reservoir Project

This document contains the summary and recommendations from the final report. A map and some photos are available on the page about the pre-study, where also the problem is presented.


The main purpose of the project has been to "accomplish a better knowledge of the natural sedimentological processes and conditions in the Arenal reservoir basin, with special attention to the gamalote problem. The results of the project should constitute a firm basis for concrete measures to overcome the gamalote problems and should be presented in a way to be useful in the planning for the future development of the region". The following is a summary of the results presented in the final report.

The report includes four introductory chapters defining the general objectives of the project, the main items of the investigation program and a brief presentation of previous subject reports. Chapter 5 follows, which is a general presentation of the Arenal drainage area. It contains data on drainage basin and river network, topographical conditions, climate, hydrology and geology, soils, land use and vegetation, intended to form a background for the discussions and analyses in the following chapters.

Chapter 6 presents the investigations in the Arenal drainage basin. It is stated that the supply of sediment to the Arenal reservoir can be studied by three main approaches:

1. Direct measurements of sediment transport loads in in-flowing river channels
2. Estimations of the release of material by soil erosion within different sub-basins
3. Measurements of the amount of deposited material in the reservoir

The first method was applied by using data from ICE hydrological stations within the drainage basins. Available ICE data on suspended sediment concentrations were supplemented by special water sampling campaigns during rainfall events in the Río Chiquito and the Río Caño Negro basins.

For the 1976/77 - 1993/94 period the estimated annual sediment load from Río Chiquito was 77,000 tonnes and from Río Caño Negro 68,000 tonnes (p.26). By extrapolation to other sub-basins in the Arenal drainage area, considering topography, rainfall, land use and other environmental characteristics, the annual sediment load for the whole Arenal drainage basin was estimated at 345,000 tonnes for the same period. Including also bed load transport the total long-term inflow of sediments to the reservoir was estimated to be in the order of 400,000 tonnes per year. The figures for Río Caño Negro may be somewhat too high and the figures for Río Chiquito somewhat too low. However, the total figure can be regarded as a safe estimation during present land use and prevailing environmental conditions.

Sediment load during most years is lower than the long-term figures given above. However, during years with extreme events the sediment load may be considerably higher. One example is the exceptional flood in Río Chiquito in July 1995, when sediment load during one single event was around 140,000 tonnes. Exceptional events with heavy soil erosion are essential for the long-term sediment load conditions. Future land use changes in the Arenal drainage basin should be planned with due attention to the risks for catastrophic events and with the intention of keeping soil erosion at a minimum.

The estimation of the release of material by soil erosion within different sub-basins has been carried through mainly by the identification of main sediment sources during field reconnaissance studies, and by using other sources of information, especially available geomorphological maps. Soil erosion models have not been applied within the project but are briefly discussed in Chapter 12. The results of bottom sediment investigations in the reservoir are presented in Chapter 9 and summarised in Chapter 12.

Chapter 7 provides some key background data on the reservoir, its operation, its morphology, and the geological setting. The wetland, Laguna Arenal, is described briefly, as a background to the investigations in the reservoir.

Chapter 8 presents the shoreline development around the Arenal reservoir during the period of operation. Two types of processes cause the main shoreline changes in the reservoir:

1. Shoreline erosion
2. Delta development

The rate of shoreline erosion is depending on several factors, of which shoreline morphology and shoreline properties on one hand, fluctuations of water level and wind directions on the other, are the most important. Due to dominating north-easterly winds and an easily eroded topsoil, shoreline erosion has so far been concentrated to the western part of the reservoir. In this part of the reservoir the topsoil is, in exposed areas, totally removed within the most frequent regulation interval (530-546 m a.s.l.). In the eastern part of the reservoir and in lee-side positions much material still remains to be eroded.

Calculations have shown that about 20% (7 million tonnes) of the totally available amount of topsoil material has been eroded during the period of reservoir operation. About 75% of the material was deposited in the useful volume, but this ratio as well as the total amount will decrease in the future and material will successively be transported to the dead storage. In the long perspective the useful volume of the reservoir will increase due to shoreline erosion, which will have a positive effect on the life length of the reservoir.

The rate of delta development is of importance for the life length of the reservoir, since the delta development mainly takes place in the useful volume of the reservoir. The amount of deposition has been investigated at the Río Chiquito delta and has been calculated at an annual maximum of 100 000 tonnes. Extrapolated to other delta areas this gives a total annual amount in the order of 300 000 tonnes deposited in the useful volume of the reservoir.

In contrast to shoreline erosion processes the future deposition of delta material will affect the life length of the reservoir. Especially the delta area of Río Chiquito is of special importance.

Chapter 9 deals with the sediment accumulated within the dead storage area of the Arenal reservoir. Sediment cores were sampled and X-rayed in order to document the vertical sequence of sedimentary structures and to determine the amount of sediment deposited on top of the pre-reservoir surface.

At several sampling stations the pre-reservoir deposits were sandy in the uppermost part and capped by gas-rich, organic layers, indicating anoxic conditions. The modern reservoir deposits mainly consist of medium silt and they are very little compacted. Some distinct sedimentary structures, that may be used as time markers, and a rhythmic variation between harder and softer layers, deposited during the rainy and the dry seasons respectively, are visible in several cores and could partly be correlated from core to core.

The sedimentation rate is higher in the eastern than in the western part of the reservoir. Especially in the former lake and wetland areas, the former Laguna Arenal, the present rate of sedimentation is very low. The mean annual sedimentation rate here amounts to only a few mm, corresponding to less than half a kg/m2. Therefore, it will probably take considerable time before the increased density caused by silt deposition will compensate for the density reduction of the old, peat bottom layers, caused by the formation and expansion of gas bubbles. The rate of sedimentation is highest in front of Río Chiquito, where a delta is formed, that with time will reach the opposite shore and thus divide the Arenal reservoir into two basins.

The mean total, annual amount of solids deposited below 530 m a.s.l., that is mainly within the dead storage area, has probably amounted to about 150,000 tonnes. About 1/6 of this amount was deposited in the reservoir to the West of the delta front area of Río Chiquito, about 2/6 within the delta front area, assumed to cover about 2 km2, and about 3/6 to the East of the delta front area of Río Chiquito.

The mean dry bulk density of the sediment deposited within the dead storage area since the creation of the reservoir is very low and amounts to 0.3. This means that the mean annual increase in sediment volume (annual sediment accumulation minus annual sediment compaction) below 530 m a.s.l. has been of the order of 500,000 m3 and that the mean annual increase in thickness of the modern reservoir deposits below this level has been about 7 mm. These figures do not include delta deposits upstream the present 530 m contour level, and they do not take into consideration the increase in volume due to up-floating "gamalotes".

The dry bulk density of the sediment deposited during the operational period of the Arenal reservoir is in general very low. The thickness, especially of the softer layers, will therefore decrease considerably with increasing depth of burial because of gravitational compaction. The present reservoir deposits below the lower storage level will probably decrease to about half its present volume (an increase in dry bulk density from 0.3 to about 0.6) during a period of 100 years, provided that the rate of sediment accumulation will be the same as at present.

Chapter 10 describes the investigations of the reservoir bottom that were performed using side scan sonar, sub-bottom profiler, echo-sounder, and diving. The analysis of the side scan sonar mosaic resulted in a map showing the size and distribution of the holes left on the bottom by up-floated islands. Those holes covered ca 12% of the peat area.

The analyses also resulted in a map of depths of the peat holes. The depths vary from 2 metres, to over 5 metres, depending on how much of the peat has floated up.

By comparing side scan sonar and sub-bottom profile data from 1990 and 1995, it could be shown that much of the up-floating has taken place after 1990. This is consistent with the observations of gamalote peat on the surface, with the largest amounts in 1994. Furthermore, the areas that were affected mostly after 1990, also showed a tendency for the holes to be deeper, i.e., the entire thickness of the peat floated up.

The side scan sonar records are lighter in a zone around the holes. This is interpreted as a sign of absence of gas. The pattern was clearer in 1990, when the water level was high and rising, than in 1995, when it was low and falling.

The sub-bottom profiles show that the peat is full of gas bubbles. In 1995 this was the case in the entire peat area. In 1990, however, the eastern half of the peat area was acoustically rather transparent in the top 2 to 3 metres, revealing that there were no gas bubbles present. The difference between the western and eastern part of the peat area was very distinct.

The data were also analysed regarding sedimentation in the reservoir. Except for the area off Río Chiquito and in the thalweg towards the Sangregado dam, only quite small amounts of sedimentation could be detected. Sediments with low organic content, i.e., those deposited quite rapidly and not too long before the survey, may be devoid of gas, and thus appear white on the sonographs and stratified on the sub-bottom profiles. On the 1995 data this was observed only in one place off Río Chiquito. On the 1990 data there were more areas with low gas content in the surface sediments.

Chapter 11 starts with a background, and then presents the main results from a thorough investigation on floating islands that was made in Scandinavia by Pousette (1965). One may note the strong significance that the methane gas has for the process. In the cold climate of Scandinavia, the temperature effect on the generation rate of methane is decisive in making them float up in the summer, although the density of the water is lower then.

On the basis of the field investigations made, e.g., when diving, an analysis of the process of peat up-floating is made, attempting to describe the complicated physical mechanisms involved. The theory explains why floating islands in the Arenal Reservoir are so "small" and numerous, compared to the conditions in many lakes with floating islands in Scandinavia.

A computer prediction model of peat up-floating was developed in the project. It uses diurnal data of the water level as input, and is calibrated against empirical "gamalote data". The model is based on the equation for methane solubility in water at different pressures, and Boyle's law, stating that the product of pressure and volume of the gas is constant (at constant temperature, which is a fair approximation within the peat). Among the unknown parameters-that have to be estimated by an iterative method-the generation rate of methane, and the vertical migration rate of the bubbles, are the most important.

The model was found to be capable of predicting the seasons when gamalote problems are likely to occur. It also reasonably correctly predicts the relative severity of the problem from year to year. However, evidently the accuracy could be improved a lot by refining the model, taking into account more of what is already known today: the structure of the peat, the fact that the generation rate of methane decreases with time, the decreasing volume of peat left in the reservoir, the accumulation of sediments on top of the peat (although this is of minor significance at present).

On the basis of indirect evidence it was deduced, that in addition, there is a close temporal correlation between wind-driven thermocline movements, and gamalote up-floating. The reason is that the density of the water increases rapidly when the cold bottom water reaches an area, leading to an equally rapid increase in buoyancy of the peat. While this effect of the thermocline appears rather likely, there may also be another one, more speculative at the present stage:

In an analysis of the possibility for peat to reach the intake under the water, the thermocline variations were hypothesised to be a main factor. Although the argumentation is purely theoretical-in the lack of adequate field observations-it provides the necessary framework for a future study of the matter.

The chapter continues with a commented summary of all proposed remedies, and ends with the conclusions from this part of the study, formulated as a number of recommendations. Those are summarised in Chapter 15.

Chapter 12 deals with the prediction of the reservoir's lifetime. It contains a discussion on total inflow of sediment to the reservoir from tributary rivers, including suspended sediment load and bed load. Possibilities to estimate soil erosion by different types of models are discussed, and the differences compared to direct data on sediment transport are pointed out. Shoreline processes and delta development, and their importance for sediment deposition in the reservoir, are also pointed out. Measurements of the deposition of fine-grained material in deeper parts of the reservoir are reported as additional parts of the data background.

After a discussion on the reliability of available data a total sediment budget is presented. The figures on average annual erosion and deposition are given in tonnes as well as in cubic meters. The budget shows that the inflow of sediment from the rivers is by far the most important source of material for reservoir sedimentation and for the reduction of reservoir capacity.

Assuming an unchanged supply of material from in-flowing rivers and from shoreline erosion in the future, it will take more than 300 years until the useful volume has been reduced by 10 % of its original capacity. During the same time the dead storage will be reduced by about 16 %, somewhat less if up-floating gamalote islands are considered.

Assuming that the sediment inflow from the rivers compared to present conditions was ten times higher, it would take about 40 years until the reservoir capacity was reduced by 10 %. Dead storage would be reduced by about 13 % during the same period. Such an alternative may be a result of intense deforestation and inappropriate land use in the future. However, these prospects must be regarded as rather hypothetical, considering the growing awareness of environmental aspects.

With progressing reservoir sedimentation the relatively narrow reservoir sections in front of the Río Chiquito delta will be filled out with sediments earlier than other parts of the reservoir. During present environmental conditions within the river basin it will take at least two centuries, until any inconvenient sediment barrier can be formed. If problems arise, measures can easily be taken to improve conditions by for instance flow diversion works, excavations or dredging.

Chapter 13 summarises briefly the teaching that was done as a part of this project, both in the form of brief courses, and as "on-the-job" training.


15.1. Introduction

On the basis of the results of the investigations we have pointed out fields, within which our present knowledge is still too weak to allow detailed conclusions and concrete remedy measures. We have also tried to define such problems in a long-term perspective, that may in the future constitute serious consequences for the Arenal area.

The problems deal both with the reservoir itself and with the surrounding drainage basin. The reservoir problems are the most critical for the present moment. This is especially true for the acute "gamalote" problem, which has to be solved by immediate measures in the intake area.

In a longer perspective the more general environmental problems are of great interest. The sedimentological aspects are significant, both regarding the sediment production by soil erosion and the sediment deposition by reservoir sedimentation. It is necessary to pay attention to these problems and to follow the future development by monitoring systems.

The following recommendations aim both at the actual and long-term problems, starting with the "gamalote" problem.

15.2. Recommendations Regarding the Gamalote Problem

The objective of any change compared to the present procedures, is either to cut costs, to increase the reliability of the power production, or both at the same time. The power plants are in operation in daytime of weekdays, but inactive during nights. However, labour costs are higher at nights and in weekends. The primary objective would therefore be to minimise the need for maintenance (inspections, cleaning, dredging, towing floating islands, etc.), while the secondary objective would be to make it possible to perform maintenance while generating. Additional comments can be found in Chapter 11, pages 101 through 106.

15.2.1. General Advises for the Operation of the Reservoir

Rapid and deep drawdowns should be avoided, and successive lowerings of the drawdown level year after year should be planned with caution.

It is recommended to calculate the "gamalote risk" by using the computer prediction model, when planning the operation of the reservoir.

15.2.2. General Advises for the Maintenance of the Intake

The prediction model is recommended to be used for forecasting the "gamalote seasons", during which a greater caution should be exercised.

During "gamalote seasons", the risk is especially high when the wind dies down after a period with a steady trade wind; cleanings and inspections are especially called for at such occasions.

A monitoring system to detect clogging of the intake should be installed before the next gamalote season starts. It could be based on monitoring the pressure difference over the intake grid during electricity generation.

Floating islands that are drifting towards the intake should be removed also in the future. It is recommended to avoid breaking the islands up. Towing them using a net might be a good solution.

15.2.3. Recommended Follow-Up Study

The results of the present study are still inconclusive concerning the best practice for decreasing the problems through mechanical constructions at the intake. However, the study has advanced the knowledge of the process to the point, that it is now possible to design an effective and conclusive follow-up study that can be completed during the next gamalote season.

The follow-up study should investigate the movements of peat under the water. Field measurements should include the near-bottom water circulation and density stratification in the intake area. Proposals for improvements of mechanical constructions at the intake should be based on the results of the study. The project should include to establish a permanent monitoring station, the data of which can be used for increasing the predicting power of the present model for forecasting gamalote problems at the intake.

The study should preferably be made during the up-coming "gamalote season". It is suggested that in parallel a study be made of the costs for implementing some of the most attractive mitigation measures, e.g., raising the intake level, or putting a net over the intake bay. In combination, this will allow a cost-benefit analysis to be made, upon which a decision on action can be based.

15.3. Recommendations Recarding Drainage Basin and Reservoir Deposits

The objective of any future policy regarding drainage area and reservoir deposits should be to minimise the sediment production in the basin and thus also reduce the rate of reservoir sedimentation. It must be possible to observe the results of future management measures by adequate monitoring systems.

The discussions and analyses in the Final Report and in the different Subject Reports have indicated several improvements that could be undertaken in existing measurement programs to facilitate monitoring and analyses. One example is the proposal to use X-ray radiographic technique in future studies of sedimentation in the Arenal reservoir (cf. Chapter 13 and "X-ray manual" published in the "Cachí Report", Axelsson 1992d).

The following measures are recommended:

1. Analyses of available hydrological data should be carried through, particularly with regard to the prediction of exceptional floods

2. Water samplings in the rivers should continue in a revised network of hydrological stations

3. Detailed land management plans for the Arenal drainage basin should be implemented

4. Recurrent re-surveying of transverse and longitudinal sections in the reservoir should be carried through, particularly in the delta areas of in-flowing rivers

A detailed program for the implementation of the recommendations, especially those considering a monitoring program, should be worked out as soon as possible.

Back to top of page