Environmental factors associated with long-term changes in chlorophyll-a concentration in the Amazon floodplain

Abstract

The study of chlorophyll-a concentration in flood pulsed wetlands has been based mostly on datasets obtained at different sites or along track lines occupied during cruises. In situ water data, however, are limited in time and space. This is a particularly serious constrain in remote regions of difficult access, such as the Brazilian Amazon floodplain waters. Moreover, in situ sampling monitoring has a high probability of undersampling. Some authors have used satellite imagery to address the wide range of spatial and temporal variability of chlorophyll-a concentration in the Brazilian Amazon floodplain. However, the authors have estimated the chlorophyll concentration in a synoptic view. Also, they don't explain the relationship between the chlorophyll concentration and other environmental parameters that might explain the reported time and space variability. Long-term environmental time series of continuously collected data are fundamental to identify and classify pulses and determining their role in aquatic systems. Based on this, this paper with the objective of analyze the chlorophyll-a concentration time series and their relationship with others environmental parameters uses in situ daily mean collected limnological (chlorophyll-a concentration, water level, water surface temperature, pH and turbidity) and meteorological (wind intensity, relative humidity and short wave radiation) through an automatic system (Integrated System for Environmental Monitoring-SIMA). SIMA is a set of hardware and software designed for data acquisition and real time monitoring of hydrological systems. The data are collected in preprogrammed time interval (1 hour) and are transmitted by satellite in quasi-real time for any user in a range of 2500 km from the acquisition point. We used Pearson correlation to determine the quantitative relation between chlorophyll time series and others environmental parameters. The periods of high INPE ePrint: sid.inpe.br/mtc-m18@80/2009/09.25.19.56 v1 2009-09-26 2 variability will be studied using the Fourier power spectrum and the time-frequency structure of chlorophyll time series will be analyzed using the wavelet power spectrum. To show the relationship between chlorophyll and the significantly time series highlighted by Pearson's correlation the cross wavelet analysis was carried out and the coherence and phase analyzed. The time series of chlorophyll-a shows two high peaks (47 µg/L and 53.30 µg/L) of concentration during a year: first during the rising water and second during the low water level. A little peak was observed during the high water level (10 µg/L). For the most part of rising, high and falling water level, the chlorophyll concentration is often low (from 2.26 µg/L to 9.11 µg/L). The causes of this were discussed. The relationship between the chlorophyll-a time series and others parameters were analyzed using the Cross Wavelet and coherence and phase concepts. With periodicities ranging from 2-60 days the chlorophyll-a concentration well agrees with turbidity and water level; and coherence ~1 and in-phase for rising and low water period. Water level dynamics and turbidity explain 68% of the chlorophyll-a time series variability.