Geographic constancy and contingency in the sensitivity of bromeliad food webs to precipitation change

SRIVASTAVA D.; ACOSTA MERCADO, D.; CAMPOS, A.; CARRIAS, J.F.; DÉZERALD O.; FARJALLA V.; MARINO N.; MACDONALD, A.; PAULA MUNHOZ DE OMENA; GUSTAVO PICCOLI; ANTIQUEIRA P.; TRZCINSKI M.K.; REALPE E.; HAMMILL E.; BARBERIS I.M.; JULIANA DA SILVA LEAL; FREIRE R.; MONTERO G.; OSPINA BAUTISTA F.; ROMERO G.

Montpellier

Congreso; 53rd Annual Meeting of the Association for Tropical Biology and Conservation; 2016

Association for Tropical Biology and Conservation

Background: Precipitation is predicted to change substantially over the Neotropics, potentially affecting the structure and functioningof freshwater ecosystems. However, we are limited in our ability to make generalisations about precipitation changes, as studies havetypically been conducted at one study site and only for shifts in mean conditions.Methods: Here we report on experimental manipulations of rainfall, replicated in 7 sites from 29°S to 18°N, and their impact on acommon ecosystem: the aquatic food webs within bromeliads. In each site, we altered the quantity of rain entering bromeliads overa two month period by 50-200%, and the dispersion of this rainfall between days by 10-300%, creating a response surface relative tosite ambient conditions. We recorded in each site a common set of macroinvertebrate, microbial and ecosystem functioning responsevariables.Results: Shifts in rainfall had general, site-specific or no effect on this aquatic system, depending on the response metric. For example,effects of rainfall on bacterial density were similar across sites. However, macroinvertebrate feeding groups often differed substantiallybetween study sites in their sensitivity to rainfall manipulations. This site contingency likely reflects geographic differences in bromeliadhydrology, litter quality and faunal traits. Despite site contingency, many sites showed optimal functioning near current climaticconditions. Finally, some ecosystem functions were largely resistant to rainfall change, such as decomposition and CO2 flux, despitethe extreme range of rainfall change.Conclusion: This study demonstrates that climate change studies based on single conditions or sites cannot be usefully extrapolatedto regional scales; only explicit inclusion of mechanisms will permit generalizations.