Open AccessModeling the vertical distribution of chlorophyll in the California Current System
Author(s) -
MillánNúñez Roberto,
AlvarezBorrego Saúl,
Trees Charles C.
Publication year - 1997
Publication title -
journal of geophysical research: oceans
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.67
H-Index - 298
eISSN - 2156-2202
pISSN - 0148-0227
DOI - 10.1029/97jc00079
Subject(s) - environmental science , chlorophyll a , empirical orthogonal functions , spatial distribution , productivity , primary productivity , photosynthetic pigment , homogeneous , water column , biomass (ecology) , atmospheric sciences , empirical modelling , current (fluid) , chlorophyll , ocean color , mathematics , oceanography , geology , statistics , physics , computer science , ecology , biology , nutrient , botany , satellite , combinatorics , astronomy , economics , macroeconomics , programming language
Remote sensing of ocean color provides data on the average photosynthetic pigment concentration in the first optical depth. To model primary productivity in the water column, estimates of the vertical distribution of pigment concentration are required. We used a Gaussian distribution function proposed by Platt et al. [1988] to represent the pigment vertical profile with four parameters. Empirical relationships were derived to estimate these parameters for the California Current System between 28° and 37°N, using California Cooperative Oceanic Fisheries Investigations (CalCOFI) data (1978–1992). The studied area was divided into six spatial subregions and temporally into warm and cool seasons. Regression models were developed for each subregion and season to estimate each of the parameters as functions of surface chlorophyll. Comparison of chlorophyll profiles obtained during the 1994 CalCOFI cruises (not used for constructing the models) with those estimated with our models shows a general agreement. Assuming a homogeneous biomass profile resulted in underestimation of integrated primary productivity (pp) by as much as 30%, whereas the modeled profiles gave equal or overestimated pp (up to 23%), with respect to the values derived from the real 1994 profiles.