Ecosystem behavior of southern Kaneohe Bay, Hawaii: A statistical and modelling approach

A rank correlation analysis of water quality parameters in the Coastal Intensive Site Network (CISNet) three-year dataset for southern Kaneohe Bay, Hawaii, was not sufficiently robust to elucidate nutrient cycling processes in the bay due to phase differences, as well as the relatively long sampling interval of the CISNet program. To quantify more fully ecosystem dynamics in southern Kaneohe Bay, the Kaneohe Bay ECOsystem Model (KECOM) was developed. This model is a nine-compartment nitrogen cycle biogeochemical box model of the pelagic and benthic systems of the southern bay. The baseline quasi-steady state of the model produces a comparable net denitrification flux to that obtained by Hoover [Hoover, D., 2002. Fluvial nitrogen and phosphorus in Hawaii: storm runoff, land use, and impacts on coastal waters. Ph.D. Dissertation. University of Hawaii at Manoa, Oceanography] for the present-day N budget of southern Kaneohe Bay. The model describing solely the N cycle captured the general trends of changes in water quality parameters in response to a particular storm perturbation as observed by Ringuet and Mackenzie [Ringuet, S., Mackenzie, F.T. Controls on nutrient and phytoplankton dynamics during normal flow and storm runoff conditions, southern Kaneohe Bay, Hawaii. Estuaries, in press]; however, it did not entirely reproduce the timing and magnitude of the biological blooms that occurred along with the bay's recovery. The results point to the possible emergence of P-limitation as a factor influencing the bay's recovery from a large storm perturbation, as suggested by observations [Ringuet, S., Mackenzie, F.T. Controls on nutrient and phytoplankton dynamics during normal flow and storm runoff conditions, southern Kaneohe Bay, Hawaii. Estuaries, in press]. The stability analysis of KECOM using a community matrix approach indicated that the instability of the benthic autotrophic biomass is caused by low grazing and/or high nutrient uptake rates due to high irradiance and potential growth rate. These two possible causes reinforce two competing hypotheses that were put forward from observational work to account for the causes of the currently observed spread of nuisance benthic macroalgae in Kaneohe Bay. Finally, the eigenanalysis of the community matrices of KECOM showed that the use of an unconventional, high power Michaelis-Menten function to describe biological assimilation and grazing kinetics was needed to stabilize the system, suggesting a future research direction involving quantification of the strength of assimilation and grazing feedbacks. In addition as anticipated, such a finding also indicates that the N cycle alone does not provide sufficient feedbacks within the Kaneohe Bay ecosystem, necessitating an ecosystem model describing all the feedbacks involving the coupled C, N and P cycles.