Abstract
Selective water release from the deeper pools of reservoirs for energy
generation alters the temperature of downstream rivers. Thermal
destabilization of downstream rivers can be detrimental to riverine
ecosystem by potentially disturbing the growth stages of various aquatic
species. To predict this impact of planned hydropower dams worldwide, we
developed, tested and implemented a framework called ‘FUture
Temperatures Using River hISTory’ (FUTURIST). The framework used
historical records of in-situ river temperatures from 107 dams in the
U.S. to train an artificial neural network (ANN) model to predict
temperature change between upstream and downstream rivers. The model was
then independently validated over multiple existing hydropower dams in
Southeast Asia. Application of the model over 216 planned dam sites
afforded the prediction of their likely thermal impacts. Results
predicted a consistent shift toward lower temperatures during summers
and higher temperatures during winters. During Jun-Aug, 80% of the
selected planned sites are likely to cool downstream rivers out of which
15% are expected to reduce temperatures by more than 6˚C. Reservoirs
that experience strong thermal stratification tend to cool severely
during warm seasons. Over the months of Dec-Feb, a relatively consistent
pattern of moderate warming was observed with a likely temperature
change varying between 1.0 to 4.5˚C. Such impacts, homogenized over
time, raise concerns for the ecological biodiversity and native species.
The presented outlook to future thermal pollution will help design
sustainable hydropower expansion plans so that the upcoming dams do not
face and cause the same problems identified with the existing ones.