Spatio-temporal dynamics of phytoplankton and primary production in Lake Tanganyika using a MODIS based bio-optical time series
Introduction
Knowledge of the spatial variations of primary production, nutrient concentration and community structure is fundamental to the understanding of ecosystem dynamics (Bootsma and Hecky, 1993). Lake Tanganyika is characterised by extensive patchiness in its water chemistry, plankton concentration and fish distribution (Plisnier et al., 1999, Salonen et al., 1999). Due to variations in incident light, mixing depth and nutrient availability, algal production is expected to be particularly variable in time and space.
The lack of extensive data on seasonal and local variations in large tropical lakes, such as Lake Tanganyika, limits the possibility to make reliable estimates of lake wide primary production. Indeed, estimates of phytoplankton biomass and primary production for Lake Tanganyika have been based on measurements made in a small number of offshore sites, often with different sampling design and estimation methods. This has led to significantly different estimates of primary production (Hecky & Fee, 1981, Sarvala et al., 1999, Descy et al., 2005, Stenuite et al., 2007). Accordingly, Descy et al. (2005) have stressed that apparent historical changes in primary production should be interpreted with care, as different methods of phytoplankton analysis could lead to different results.
To improve the assessment of primary production and its dynamics in Lake Tanganyika, it is necessary to extend information gained from spatially and temporally localised measurements to the entire lake. On the other hand, the synoptic perspective provided by satellite data provides valuable information regarding the surface spatio-temporal characteristics of extensive ecosystems. The use of backscattered solar radiation to model global distributions of phytoplankton primary production has provided valuable insights into ocean dynamics and global biogeochemistry (Bricaud et al., 2002, Tilstone et al., 2009). However, little has been done to develop approaches appropriate for lake ecosystems, in particular in the southern hemisphere.
In the present study, we used remotely sensed estimates of phytoplankton biomass (chlorophyll-a) and optical properties (K490) to perform a regionalisation of Lake Tanganyika according to geographic areas with similar temporal patterns. From these estimates, regional patterns of daily primary production were calculated, using photosynthesis parameters derived from in situ 14C incubations. Finally, we calculated the overall lake phytoplankton productivity and compared it with other studies.
Section snippets
Data and methods
Level 1B MODIS-AQUA images (1 km2 resolution) were selected from July, 2002 to November, 2005. The parameters of interest, namely chlorophyll-a concentration (chl-a) and the diffuse attenuation coefficient at 490 nm (K490), were derived and optimized on the basis of in situ data and considering local atmospheric conditions (see Horion et al., 2010-this volume). The optimization utilised data from the CLIMLAKE cruises sampling series (Descy et al., 2006).
The missing values (pixels) in the daily
Spatial analysis
The EOF analysis of the extensive satellite dataset showed a variance dominated by the first 11 modes for the chlorophyll-a and by the first 6 for the light attenuation coefficient. The proportion of the spatio-temporal variance explained by these EOFs is larger than 94% (Table 1).
The spatial regionalisation of the Lake Tanganyika, based on EOFs modes, showed a similar spatial structure for both chl-a and K490, with a higher variability present in the time series of chl-a images. In particular,
Discussion
An extensive dataset of satellite images provides an opportunity to track dynamics in ecosystems with high spatial and temporal variabilities (Alvera-Azcárate et al., 2005). For Lake Tanganyika, a filled spatio-temporal matrix for chlorophyll-a and K490 was built at a spatial resolution of 1 km2 (lake area = 31,745 pixels) for an average of 394 days, extending from 5/7/2002 to 30/11/2005. The analysis of this extensive dataset confirmed the patterns observed in the field studies and allowed us to
Conclusions
Large lakes may be a good metric for global processes: large enough to damp out localised phenomena, yet small enough (compared to oceans) to generate observable responses. The African Great Lakes are sensitive to climate change and therefore serve as important long-term indicators. Previous research has established that Lake Tanganyika may be affected by global change through a decrease in primary production. The present study demonstrates that satellite based measurements can provide valuable
Acknowledgements
The CLIMLAKE and CLIMFISH projects were financed by the BELSPO (Belgian Science Policy Office), in partnerships with TAFIRI (Tanzanian Fisheries and Research Institute) and the Department of Fisheries (Zambia). We thank the Belgian Federal Science Policy Office and the Belgian Technical Cooperation for co-financing CLIMLAKE through a frame agreement with the Royal Museum for Central Africa (Tervuren, Belgium). Nadia Bergamino had a fellowship from the University of Siena PAR funds for
References (35)
- et al.
Reconstruction of incomplete oceanographic data sets using Empirical Orthogonal Functions. Application to the Adriatic Sea surface temperature
Ocean Modelling
(2005) - et al.
Examining the dynamics of phytoplankton biomass in Lake Tanganyika using empirical orthogonal functions
Ecological Modelling
(2007) - et al.
Algal biomass and sea surface temperature in the Mediterranean Basin: Intercomparison of data from various satellite sensors, and implications for primary production estimates
Remote Sensing of the Environment
(2002) A comparison of global estimates of marine primary production from ocean color
Deep-Sea Research II
(2006)- et al.
Spatio-temporal dynamics of chlorophyll-a concentration and attenuation coefficient K490 in Lake Tanganyika, Africa. Part 1: Regional calibration of daily bio-optical time series retrieved from MODIS/Aqua imagery
Remote sensing of Environment
(2010) - et al.
Measured and remotely sensed estimates of primary production in the Atlantic Ocean from 1998 to 2005
Deep Sea Research Part II: Topical Studies in Oceanography
(2009) - et al.
EOF calculations and data filling from incomplete oceanographic datasets
Journal of Atmospheric and Oceanic Technology
(2003) - et al.
DINEOF reconstruction of clouded images including error maps — Application to the sea-surface temperature around Corsican Island
Ocean Science
(2006) - et al.
Conservation of the African Great Lakes: A limnological perspective
Conservation Biology (Special Issue)
(1993) - et al.
Optimal analysis of in situ data in the western Mediterranean using statistics and cross validation
Journal of Atmospheric and Oceanic Technology
(1996)
Late Holocene linkages between decade-century scale climate variability and productivity at Lake Tanganyika, Africa
Journal of Paleolimnology
Lake Tanganyika and its life
Phytoplankton pigments and community composition in Lake Tanganyika
Freshwater Biology
Climate variability as recorded in Lake Tanganyika (Climlake): Final report. Belgian Science Policy
Primary production and rates of algal growth in Lake Tanganyika
Limnology and Oceanography
The phytoplankton and protozooplankton of the euphotic zone of Lake Tanganyika: Species composition, biomass, chlorophyll content, and spatio-temporal distribution
Limnology and Oceanography
Phytoplankton ecology of the great lakes in the rift valleys of Central Africa
Archiv für Hydrobiologie, Beiheft Ergebnisse der Limnologie
Cited by (70)
Remote sensing-based water quality monitoring in African reservoirs, potential and limitations of sensors and algorithms: A systematic review
2024, Physics and Chemistry of the EarthPhytoplankton blooms and fish kills in Lake Tanganyika related to upwelling and the limnological cycle
2023, Journal of Great Lakes ResearchLake Tanganyika: Status, challenges, and opportunities for research collaborations
2023, Journal of Great Lakes ResearchSpatiotemporal patterns and drivers of phytoplankton primary productivity in China's lakes and reservoirs at a national scale
2023, Global and Planetary ChangeEvolving phytoplankton primary productivity patterns in typical Tibetan Plateau lake systems and associated driving mechanisms since the 2000s
2022, Remote Sensing Applications: Society and EnvironmentRemote estimation of phytoplankton primary production in clear to turbid waters by integrating a semi-analytical model with a machine learning algorithm
2022, Remote Sensing of EnvironmentCitation Excerpt :An alternative approach is to use in situ information to directly assign photosynthetic parameters on a pixel-by-pixel basis with the aid of historical databases (Bouman et al., 2018; Kulk et al., 2020). This kind of approach assumes that averaged PmaxB and Ek are representative enough in a distinct region and a specific season, which has been operationally used to estimate IPP in the global ocean (Kulk et al., 2020; Longhurst et al., 1995; Platt and Sathyendranath, 1999) and in Lake Tanganyika (Africa) (Bergamino et al., 2010). In addition, a more dynamic strategy for the assignment has been proposed to use a nearest-neighbor method that considers the ambient characteristics of Chla and sea surface temperature (SST) (Platt et al., 2008).