Detecting intra- and inter-annual variability in gross primary productivity of a North American grassland using MODIS MAIAC data

Prairie productivity is largely affected by temperature and precipitation and is vulnerable to a changing climate. In this study, we used 5-years of growing season eddy covariance and satellite data (MODIS MAIAC) from two adjacent field sites in southern Alberta (Canada) prairie to monitor intra- and inter-annual variation in prairie productivity using remote sensing. Three MODIS vegetation indices were examined to track seasonal variation of Gross Primary Productivity (GPP): the normalized difference vegetation index (NDVI), the NIRv, and the chlorophyll/carotenoid index (CCI). The productivity of these prairie ecosystems was mainly driven by precipitation, with temperature affecting the starting time of the growing season. The three vegetation indices captured distinct aspects of GPP phenology. CCI, which is sensitive to chlorophyll and carotenoid pigment ratios, followed seasonal GPP dynamics more closely than NDVI and NIRv. Consequently, less hysteresis occurred with the seasonal CCI-GPP relationship than with the NDVI-GPP or NIRv-GPP relationships. Relative to 16-day composite data, daily MODIS data provided a more detailed GPP phenology. However, relationships between all vegetation indices and GPP improved with temporal aggregation up to one month, demonstrating that the degree of data aggregation affects the ability of reflectance-based indices to track GPP. Results from a multivariable regression revealed a strong relationship between GPP and a linear combination of 3 MODIS bands (B1, B2 and B11), which indicates that additional spectral information provided by the MODIS ocean band (band 11) can help track grassland GPP better than typical 2-band broad-band indices (e.g. NDVI or NIRv) only. Improved monitoring of prairie ecosystems using these enhanced approaches, can lead to a better understanding of the effects of changing weather and climate on the productivity of prairie ecosystems.