Factors Controlling Respiration Rates and Respired Carbon Dioxide Signatures in Riverine Ecosystems of the Amazon Basin

This study examined the processes controlling respiration rates observed in streams and rivers throughout the Amazon basin during the dry season by substituting spatial coverage for experimental manipulation. Throughout the Brazilian states of Amazonas and Acre, respiration rates ranged from 0.066 to 1.45 μM/hr of O2 consumed. In situ respiration was positively correlated with pH (r2=0.60), with pH values ranging from 3.95 to 8.57. Although the concentration of bulk size fractions of organic matter(dissolved organic carbon (DOC), fine particulate organic carbon, and coarse particulate organic carbon) were uncorrelated with both pH and respiration, respiration was positively correlated with the percentage of DOC that was less than 5 kDa as determined by centrifuge ultrafiltration (r2=0.52). No correlation was observed for the less than 100 kDa fraction. Further, pH was also correlated with the percentage of DOC in the <5 kDa fraction (r2=0.86), as the <5 kDa fraction increased from 34% in acidic blackwater streams to 91% in more basic whitewater rivers. These results suggest that low molecular weight organic matter (LMWOM, <5 kDa) is labile and supports higher respiration rates as compared to high molecular weight organic matter, and that pH may control the size distribution of dissolved organic matter. Further, at high pH sites with high respiration rates, net primary production ranged from 3.54 to 13.5 μM/hr of O2 produced. These rates suggest that higher pH sites are dominated by in situ production, resulting in high yields of LMWOM, which is rapidly consumed during the dry season. The 13C of respired CO2 was monitored during bottle incubations to characterize the source of organic matter being respired. Values ranged from -15.2 to -27.0‰, similar to the 13C of DIC at each site, indicating that respiration is a key process controlling the δ13C of the DIC. Furthermore, there is a positive correlation between the δ13C of respired CO2 and respiration rate (r2=0.76), suggesting that the fuel for respiration is either C4 vegetation or an isotopically enriched pool of C3 based organic matter that is selectively respired. Given that the <5 kDa pool of organic matter is positively correlated with respiration, we hypothesize that this pool may be isotopically enriched and derived from autochthonous material, leading to a heavy isotopic signature of the respired CO2. Autochthonous carbon production may play a larger role in riverine carbon cycling in the Amazon than previously thought during the dry season.