El Quimbo, the second large hydropower dam built across the river Magdalena in Colombia, began operating in late 2015, and has already generated more headlines than power, for discharging poor quality water downstream. Its design had overlooked serious technical/ environmental problems that are common to large tropical reservoirs and the consulting firm who prepared the Environmental Impact Assessment (EIA) had committed grave errors in analyzing those problems. The government oversight agency did not detect the problems and allowed the construction to go ahead, probably because of technical incompetence. However, corrupt practices cannot be ruled out either, as hundreds of million dollars are at stake in this first-ever, design-build-operate hydropower concession.

In 1987, the Colombian government built the first dam across the river Magdalena, the country's most important river system in terms of transport, fishing and agriculture. That reservoir, built at Betania, some 30km upriver from Neiva, departmental capital of Huila, was supposed to produce energy for the national grid (capacity - 540MW) and control the river's devastating floods.  However, the reservoir's relatively small active storage (850 Mm3) compared to the river flow (vary from 100 to 2000 m3/s with 475 m3/s as the annual mean) required it to be managed at an optimum level to fulfill those goals. Yet, the publicly administered power project never got close to achieving them in a satisfactory manner.

In 1996, government used that failure as a pretext, and handed over its management to the multinational energy firm EMGESA, whose only obligation was to produce energy. Private aquaculture activities began in the reservoir around 1994, and from 2005, caged tilapia population there increased exponentially, under authorizations given by the departmental government. A consulting firm estimated the limiting fish capacity of Betania to be 30,000 tons, but nobody controlled those limits. The concession did not oblige EMGESA to care for fish farming in the reservoir and its water use schedule for power generation constantly clashed with the needs of the aquaculture industry. In 2007, a severe reduction of river inflow to Betania, in combination with EMGESA's need to maximize power generation, suddenly exposed the caged fish to reservoir bottom-waters with little oxygen, killing them at unprecedented levels. However, to this day, these conflicting interests on the use of Betania water have not been resolved. 

In 2008, the Colombian government opened a concession to build and operate another hydropower dam in Magdalena, up the river from Betania, and EMGESA won it. 
EMGESA pamphlet - June 2011
Immediately, it presented plans for a dam at El Quimbo, some 14km above Betania reservoir and 1.3km above the confluence of river Paez, a major tributary. This new reservoir has the capacity to retain 9 months of Magdalena river flow and to produce 400 MW of energy before discharging that water immediately downstream. These characteristics allows EMGESA to partially control Magdalena's floods and stabilize the inflow to Betania, to smoothen its energy production fluctuations. This proposal looked so good in theory, that even the aqua-culturists in Betania supported it with enthusiasm. 

However, such a deep (close to 150m) reservoir in the tropics that floods a long, narrow canyon could cause serious environmental problems, as well as initiate grave social conflicts, heavily elevating the costs of construction, operation and mitigation, to the point of making the project no viable.

In this article, we will analyze only the possible environmental problems of El Quimbo reservoir, using publicly available information, leaving aside the socioeconomic issues, which have already generated a lot of heated discussions and protests (www.quimbo.com.co)

Design change by EMGESA after receiving the license to construct
The consulting firm INGETEC prepared the EIA for the project in 2008 (ref 2), which served as the base document to obtain the license to build the reservoir. According to our analysis, that document contains several technical errors, which, if detected in time, would have forced substantial changes to the dam design to make it environmentally viable. The government authority on the matter, ANLA, apparently, did not notice those errors and approved the project design without any changes.
Subsequently, EMGESA changed at least one crucial parameter of the dam design, which in our opinion, increased greatly the economic benefits to the company and affected the quality of the water discharged by the turbines. Such a change needed an exhaustive revision of the construction license, but the relevant authorities kept silent.

Information published by EMGESA on 29th Oct 2008 (ref 1):
According to this data, the estimated average inflow to the El Quimbo reservoir is 243 m3/s. The maximum operational water level (reservoir full) is at 720m above mean sea level and the minimum operating level, at 690m (when the reservoir draws down to this level, turbines are shut down). The active storage of the reservoir (volume between the maximum and minimum operating levels) is 1824 Mm3 and its dead storage (volume of water below the minimum operating level, hence not useful for generating power), 1381 Mm3. This data coincides with the values used in the EIA (ref 2).

Information published by EMGESA on June 2011 (ref 3):
The estimated average inflow to the reservoir is now at 237 m3/s. The maximum operational water level remains the same, at 720m, but the minimum level is now reduced to 675m (15m below that of 2008 data). With that change, the active storage volume jumped up to 2354 Mm3 (30% increase) and the dead storage decreased by the same percentage to 861 Mm3.

No publicly available document explains why such a change in minimum operating water level was made in El Quimbo reservoir and how it was done. The contractor modified the parameter after receiving the license to build, but we could not find a document authorizing such a change. We can only guess the reasoning behind the change, as it greatly increases EMGESA's capacity to generate power, and hence its profit. Below, we will explain how this modification could substantially lower the quality of water discharged by the turbines, needing a revision of the previous authorization.

Reservoir minimum operating level and turbine discharge water quality
Downstream water users of an artificial reservoir -aquatic life, animals and humans- all depend on the quality and the quantity of water discharged from the dam. During the first filling of the hydropower dam, a limited flow is released downstream through the reservoir bottom discharge gates until sufficient water is accumulated in the reservoir to begin operating the turbines. From then on, the quality of the downstream water flow depends on the quality of the turbine discharge. The bottom discharge gates are opened only intermittently, to remove bottom sediments or when the turbines are being repaired.

If the river does not bring contaminated water to a reservoir, downstream water quality depends primarily on water temperature and its dissolved oxygen content at the level water enters the discharge mechanism (bottom gates or turbines). As we go down to the bottom water layers of a reservoir, water temperature decreases, since the number of solar rays penetrating deeper reduces gradually. Cooler water is denser and thus stays at the bottom of the reservoir. In shallower areas of a lake, crosswinds help mix the bottom and topmost waters. In deeper parts, lake water always forms layers of decreasing temperature as one goes down.  

The concentration of dissolved oxygen in lake water decreases as the reservoir's aquatic life consumes it. If the reservoir bottom was not cleared of its original vegetation and/or if the river brings in heavy organic matter to the reservoir, their decaying process also consumes a lot of oxygen in the bottom layers, until it is completely depleted. When water does not mix between the layers formed by temperature-based stratification, bottom layers of deep lakes always remain cooler and devoid of oxygen. If there are a lot more vegetative matter in the bottom, decaying occurs through anaerobic processes, producing gases like methane and hydrogen sulfide, giving rise to foul-smelling lakes.

In reservoirs located in latitudes closer to the poles, during the fall season, surface water begins cooling faster than the deeper layers. Since water density reaches a maximum at 4oC, as the fall turns into winter, this oxygen-rich dense water travels deep down, mixing and replenishing the lake bottom waters with oxygen, once a year at least. However, in tropical reservoirs like El Quimbo, such a thermal inversion of waters does not take place and the bottom waters always remain cold and with little oxygen.

In the EIA, section Captación y conducción mentions that the trash rack at each turbine intake is seated at the level of 655m, thus, the minimum temperature and dissolved oxygen content of water entering a turbine correspond to that level. When the reservoir is at its maximum operational level (720m), the bottom-most water entering the turbines is 65m deep. According to the 2008 dam design (ref 1), even during low river flow periods, the reservoir level would not go below 690m, meaning water to the turbines would come from layers at or above 35m depth. Given that the trash racks are 15m high, when the reservoir is at its lowest, top of the turbine intake will still be 20m below the surface water level. It is common to keep a certain minimum head of water above the turbine intake so that air does not enter the turbines and damage them. Nevertheless, in our opinion, a value of 20m is too much, and the turbine intake could have been raised a few meters more, so that the turbines can use water from higher layers, thus of a better quality. However, the EIA does not make any comment on this issue.

Now, as per EMGESA's dam design of 2011, the minimum operational water level is 15m lower. This leaves only a 5m minimum height of water above the turbine intake, assuming EMGESA did not change its location also. If that assumption were correct, when the reservoir is at its minimum operational level, the quality of turbine discharge water would be better than the previous design. However, if EMGESA had lowered the level of the turbine intake (such information is not available publicly) then the water quality of the turbine discharge would mostly be worse than the previous situation. In any case, EMGESA must have changed the design to increase its profit margin, without caring much about the quality of turbine discharge.

The EIA mentions that El Quimbo dam has a bottom discharge gate at 605m, so that, during the first filling of the reservoir, a minimum of 36 m3/s of flow can be discharged downstream, to safeguard the ecosystem immediately below the dam. At the beginning, this discharge would be of acceptable quality, since the reservoir water level would still not be too deep. However, as the water depth behind the dam increases, its quality will deteriorate. When the turbines begin functioning, the bottom discharge gate should be shut down.

Doubts about El Quimbo reservoir water quality predictions
To model the water quality for the EIA, the consulting firm applied the software package WQRSS3 in three different situations: within El Quimbo reservoir; in the river stretch between the reservoirs El Quimbo and Betania; and within Betania reservoir. Nevertheless, this software, developed by the U.S. Army Corps of Engineers in the 1980s, has severe limitations in accurately modelling such water bodies with contrasting characteristics. The software models the temperature distribution of a water body, dividing its depth into small layers of water and then calculate the depth-wise distribution of different water quality parameters, like dissolved oxygen. However, it cannot model the mixing of water between different layers in a reservoir (ref 5). As we discussed earlier, in tropical reservoirs, thermal-inversion-related mixing of surface and deep water layers does not occur, but there could be other ways, depending on specific characteristics of each reservoir. More mixing is expected when a reservoir has: 1) a width large enough for winds to stir-up surface waves; 2) a short hydraulic retention time (time to fill the total storage of the reservoir under average river inflow); and 3) a short distance between the dam and the tail end of the reservoir (where the river enters it).  

Source: EMGESA - 2008
The EIA states that El Quimbo reservoir has a width of 1.4km, averaged over the full length of 55km between the dam and the tail, and the dominant wind direction is almost perpendicular to the canyon axis. Its hydraulic retention time is about 259 days. Under these characteristics, we would expect very little mixing between deep and surface waters to occur in El Quimbo reservoir, and the software used might be able to model this reservoir accurately.

Nevertheless, the water quality modelling results for El Quimbo, presented in the EIA, seem unrealistic. At the end of first year of filling the reservoir, EIA predicts average water temperature at the surface to be 27oC and at a depth of 65m (turbine intake), 23oC. The dissolved oxygen varies from 6 mg/l at the surface to 4 mg/l at the depth of turbine intake (a value below 2 mg/l is considered critical for aquatic life). Given the characteristics, we expected a much lower oxygen concentration at such a depth at El Quimbo. Thus, we suspected errors in model calibration parameters.

In effect, the EIA says the consultant calibrated the model using the oxygen critical situation that occurred in Betania reservoir in 2007. The software model, when applied to that situation, produced a 3mg/l dissolved oxygen value near the surface of Betania. The consultant argues that the tilapia fish in Betania, trapped in densely populated cages, would have suffocated from lack of oxygen, resulting in extremely high mortality rates. Examining the dissolved oxygen chart presented by EIA in that case, we noted that the oxygen level predictions in Betania increase at greater depths! Probably, the consultant adjusted the model parameters too much, to reproduce the low oxygen situation near the surface, resulting-in distortions elsewhere.

In our opinion, this software cannot satisfactorily model the water quality in Betania reservoir, which has characteristics very favorable to mixing of water layers. It is much wider and much shorter than that of El Quimbo, with a hydraulic retention time closer to 30 days. The EIA consultant erred in calibrating the model parameters to suit the Betania reservoir and then applying it to El Quimbo reservoir.

Common sense would have signaled the consultant to re-examine his model-predicted oxygen values for El Quimbo. If Betania reservoir, with conditions so favorable to mixing of waters, would produce critical oxygen conditions at water layers as shallow as 12m, how could the reservoir of El Quimbo, where no mixing of waters is possible, produce higher oxygen levels at depths five times greater?

The EIA goes on to compare El Quimbo with several other existing Colombian reservoirs, but the comparison lacks information on how the water quality behaved in conditions similar to those of El Quimbo. Experiences are abundant in literature (ref 4) where many reservoirs with depths shallower than that of El Quimbo had discharged very poor quality water, destroying all aquatic life downstream. That is why an agency like U.S. Tennessee Valley Authority, which manages dozens of large hydropower dams, is now implementing various mechanisms in its reservoirs to oxygenate the turbine discharges (ref 6).

Given the characteristics of El Quimbo reservoir, we expect its turbine discharges to have low temperatures and little dissolved oxygen in the long-term, causing serious environmental impacts downstream. However, the EIA discards such possibility, based on its flawed modelling studies.

Growth of aquatic plants in El Quimbo reservoir
Betania reservoir with little aquatic weeds
The EIA claims (ref 2 –p7.2.52) El Quimbo reservoir would not suffer from rapid spreading of aquatic weeds, like water hyacinth, citing the experience in Betania reservoir. Nevertheless, Betania's situation is quite different: its very low hydraulic retention time and the high discharge capacity of its turbines make the reservoir water level rise and fall quite rapidly, not allowing the weeds to take root along the lake edge. Besides, its open water surface permits wind-aided waves that quickly disperse the weeds. El Quimbo has opposing characteristics as we discussed before, and in addition, it will receive greater nutrient loads from the fertile agricultural area along the river upstream. Therefore, EIA's conclusion is unacceptable.

Aquatic weeds proliferate in La Esperanza reservoir
In Ecuador, large reservoirs under situations similar to El Quimbo (narrow lake width and high retention time) like Daule-Peripa and La Esperanza, continue to suffer from aquatic weed proliferation for over 20 years.

Aquatic weeds complicate lake navigation and are a nuisance to downstream water users. Moreover, they could initiate serious health hazards, breeding mosquitos that transmit various diseases, like Dengue, Malaria, Chikungunya and Zika. Many large tropical reservoirs all over the world have caused such health hazards (ref 4) and we expect similar threats from El Quimbo too. However, the EIA minimizes such risk, based on its erroneous conclusion above.

Conclusions and recommendations
Our evaluation has detected serious environmental hazards from El Quimbo project, which were ignored by the government agencies, probably because of technical incompetence or because they suffer from corruption-blindness. These problems are long-term and will aggravate as time goes-by, thus cannot be mitigated by patchy repairs or dolling out money. 

When you add these environmental problems to the already snowballing socioeconomic damages caused by the project, we believe that, to prevent a major disaster from taking place, it is not too late, nor too costly, to dismantle the dam and let the river flow free. Given the serious violations to the due process committed by the private entrepreneur who sponsor the dam, the Colombian government treasury will not have to pay a cent and the contractor is obliged to restore the location to its original form at his expense.

If the Colombian government is seriously worried about the lack of hydro-energy for local consumption, it can solve the problem by spending less than 10% of what it designates for dam construction: by spending that on storing water in the soil (ref 7, p95-106), in upper catchment areas of every river.  The solution is simple: in agricultural lands, help the farmer with erosion prevention barriers, furrows, trenches and terraces; in deforested lands, restore the vegetative cover with native plants; in eroded rivulets and streams, install check dams and plant bank-stabilizing trees. The water retained in upper catchments through these mechanisms will slowly drain down to the rivers, maintaining a stable flow even during the drought seasons, so that existing hydropower dams can supply constant amounts of power throughout the year. As an extra, the farmers will see their productivity increase and the restored forest cover will reduce negative effects of Climate Change too.

PS: Now, in 2016, with only a month after re-initiating the turbine operations in El Quimbo, the fisher folk downstream and in Betania are on alert because of the low quality water being discharged. The worse is yet to come, however, when the vegetation flooded by the reservoir begins to decompose.

Kashyapa A. S. Yapa, Ph.D. in Civil Eng. (UC Berkeley)
Riobamba, Ecuador.
October 2013/ re-edited and translated for the public in January 2016.

1. EMGESA (2008) “Proyecto hidroeléctrico El Quimbo”, XIV Congreso del MEM, Cali, Colombia.
2INGETEC S.A. (2008) “Estudio de impacto ambiental del proyecto hidroeléctrico El Quimbo”,          presented to EMGESA and published by Ministry of Environment, Colombia.
3. EMGESA (2011) “Proyecto hidroeléctrico El Quimbo”, promotional leaflet, june.
4.   World commission on dams (2000) “Dams and development: a new framework for decision-making”, Earthscan Publications Ltd, London and Sterling, VA.1
5. Wells, S.A. “A 2-D hydrodynamic and water quality model, CE-QUAL-W2 Version 3” http://www.ce.pdx.edu/w2/download_model_information.html
7.  Yapa, Kashyapa, A.S. (2013) "Prácticas ancestrales de crianza de agua", Min. Ambiente/ Sec.     Nacional de Gestión de Riesgos/ PNUD, Quito, Ecuador.      https://www.academia.edu/4466302/PRACTICAS_ANCESTRALES_DE_CRIANZA_DE_AGUA_-jul_2013

If you would like to comment or ask a question about this article, please write to me.

Back to INDEX