Spatial and Temporal Responses of NDVI to Climate and Soil Factors in the Grassland-forest Transition Zone of Saskatchewan, Canada

Abstract

The Normalized Difference Vegetation Index (NDVI) is generally recognized as a reliable indicator of terrestrial vegetation productivity. Understanding climatic influences on NDVI enables the prediction of productivity changes under different climatic scenarios. In this paper, we examine the role that vegetation productivity plays on both climate and NDVI through a joint analysis of temperature, precipitation, drought, and soils. Our focus is on the prairie grassland to boreal forest transition zone in Saskatchewan Canada where understanding vegetation productivity change is vital to the sustainability of various ecosystems. We analyzed 21-years (1981-2001) of monthly growing season (May-October) NDVI values derived from the National Oceanic and Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR) sensor. Monthly temperature and precipitation data were interpolated from 117 weather stations inside and around the borders of the study region. The modified Palmer Drought Severity Index (PDSI) was employed to indicate drought severity, and its monthly and growing season values were calculated based on interpolated climate factors and soil conditions. We found that significant temporal and spatial variations exist in the correlations between NDVI and temperature, precipitation, and modified PDSI. Temperature is only highly correlated with NDVI in some parts of the study area in May, and no apparent indication of this correlation is shown in other seasons. Precipitation is correlated with NDVI during the growing season in different regions of the study area, with moderate to strong correlations in June and October. From July to September, NDVI is highly correlated with PDSI in most parts of the study area. When the growing season is examined as a whole, the results show that NDVI is highly correlated with precipitation in the northern area of the study region and with PDSI in southern parts of the study region, while only small parts of the study region show a highly correlation between NDVI and temperature.