Stable isotope data as reef food-web descriptors in a dynamic tropical environment

Abstract

Despite the increasing use of stable isotope data as diet and trophic position descriptors in food-web ecology, their generic value relies on basic assumptions of constant trophic step enrichment, steady state conditions and accuracy of isotopic analyses for the estimation of trophic level (TL). This thesis explores the implications of these assumptions for understanding reef community trophodynamics in Oman, where upwelling events produce seasonal patterns among potential food sources. Nitrogen isotopes (δ15N) revealed 3-4 TL’s (6.16 to 17.8‰) and the wide range of carbon isotope (δ13C) values (-21.92 to -6.43‰) indicated that there were both benthic and pelagic sources of production. Primary producer and primary consumer δ15N fluctuated seasonally by 2.24‰ (9.02-11.26‰) however, this variability was not consistently observed at higher TLs, and therefore the δ15N of high-TL consumers may not accurately reflect their trophic position. Long-lived marine bivalves had no temporal isotopic variability, allowing the trophic position of higher consumers to be estimated using their δ15N as a baseline. Baseline organisms also allowed spatial comparison of part of the trophic structure of two reef communities; Barr al Hickman had δ15N values 2.7‰ enriched compared to a similar community at Bandar Kayran 360km to the north. Across the Western Indian Ocean, macroalgae δ15N values correlated well with differences in underlying nutrient regimes of surrounding waters. Trophic-step fractionation in herbivorous fish was 4.69-5.25‰, higher than the generally accepted value of 3.4‰ and was explained by a dynamic model incorporating feeding rate, diet quality and excretion rate, which are inherently different between herbivorous and carnivorous fish. δ15N was strongly correlated with body size in some fish species but across the entire community body size was a poor descriptor of trophic position. Use of dietspecific trophic-step fractionation values and sulphur isotopes (δ34S) greatly improved the resolution of food-web models.