Indirect upstream effects of dams: consequences of migratory consumer extirpation in Puerto Rico



Short name: 


Data set ID: 



Large dams degrade the integrity of a wide variety of ecosystems, yet direct downstream effects of dams have received the most attention from ecosystem managers and researchers. We investigated indirect upstream effects of dams resulting from decimation of migratory freshwater shrimp and fish populations in Puerto Rico, USA, in both high- and low-gradient streams. In high-gradient streams above large dams, native shrimps and fishes were extremely rare, whereas similar sites without large dams had high abundances of native consumers. Losses of native fauna above dams dramatically altered their basal food resources and assemblages of invertebrate competitors and prey. Compared to pools in high-gradient streams with no large dams, pool epilithon above dams had 9 times more algal biomass, 20 times more fine benthic organic matter (FBOM), 65 times more fine benthic inorganic matter (FBIM), 28 times more carbon (C), 19 times more nitrogen (N), and 4 times more non-decapod invertebrate biomass. High-gradient riffles upstream from large dams had 5 times more FBIM than did undammed riffles but showed no difference in algal abundance, FBOM, or non-decapod invertebrate biomass. For epilithon of low-gradient streams, differences in basal resources between pools above large dams vs. without large dams were considerably smaller in magnitude than those observed for pools in high-gradient sites. These results match previous stream experiments in which the strength of native shrimp and fish effects increased with stream gradient. Our results demonstrate that dams can indirectly affect upstream free-flowing reaches by eliminating strong top-down effects of consumers. Migratory omnivorous shrimps and fishes occur throughout the tropics, and the consequences of their declines upstream from many tropical dams are likely to be similar to those in Puerto Rico. Thus, ecological effects of migratory fauna loss upstream from dams encompass a wider variety of species interactions and biomes than the bottom-up effects (i.e., elimination of salmonid nutrient subsidies) recognized for northern temperate systems.
Date Range: 
2008-06-01 00:00:00 to 2008-06-03 00:00:00

Publication Date: 

2011-03-16 00:00:00



Additional Project roles: 

Name: Miguel C Leon Role: Data Manager
Name: Catherine Pringle Role: Associated Researcher
Name: Jeff G. Holmquist Role: Associated Researcher
Name: William H. McDowell Role: Associated Researcher


See Greathouse et al. (2006) and Greathouse (2005). (full reference in the "REFERENCES" section below). Our main data set compares dammed and undammed streams that are high gradient (i.e., above waterfalls that block migration of predatory fishes). We used a natural experiment approach (sensu, Diamond 1986), sampling macroconsumers, benthic resources, and non-decapod invertebrates in seven stream reaches upstream from large dams and 10 undammed reaches (Fig. 1a). We refer to sites above large dams as ''dammed'' and sites with no large dams as ''undammed''; however, low-head dams (height ,15 m) may occur upstream or downstream from any of our sites. The 10 undammed sites were dominated by shrimp and Sicydium with no predatory fishes. Physical conditions of the seven dammed sites are characteristic of streams that are dominated by shrimps and Sicydium (Table 1); it is highly likely that they were dominated by shrimp and Sicydium prior to dam construction. High-gradient dammed sites were above large dams without regular ''free crest'' spillway discharge (i.e., water flowing over the sloping face of a dam; hereafter referred to as spillway discharge). A lack of regular spillway discharge was previously found to cause upstream extirpation of shrimps and native fishes (Holmquist et al. 1998). Thus, our highgradient stream survey was designed to compare sites where native shrimps and Sicydium were extirpated (dammed) and sites with relatively natural shrimp and Sicydium assemblages (undammed). We chose sites so that physical and chemical parameters would not confound our comparison of dammed vs. undammed rivers.



Time Period: