Phenologic metrics derived from MODIS NDVI as indicators for Plant Available Water-holding Capacity

Soil, an essential component of agricultural ecosystems, has high spatial variability. Plant growth reflects this variability in complex interactions with other factors such as rainfall and temperature. In the Mediterranean-type dryland cropping region of South Australia, water is the main driver of crop growth variability. Plant Available Water-holding Capacity of soil (PAWC) is the soil property that measures the maximum amount of plant extractable water which can be held in the soil. It interacts with weather conditions, governing crop growth. Thus understanding spatial variability of PAWC is crucial for farm and regional agricultural management. However, the physical measurement of PAWC is costly and time consuming. Crop phenology, the timing of plant growth and development, is influenced by climatic, soil, and management factors. We hypothesised that by keeping management and climate constant by focusing on a small geographic area, the dynamic response of plants is largely due to soil functions. The objective was to use remote sensing derived phenology to understand the complex climate soil interactions. We compared phenologic metrics of annual winter-growing crops from adjacent high and low PAWC soils at two farms (Whariminda and Minnipa) in the Eyre Peninsula, South Australia. The phenologic metrics were derived for 13 seasons from time series of Moderate Resolution Imaging Spectroradiometer (MODIS) NDVI data, at 250 m spatial resolution and 16 days temporal resolution. Wilcoxon signed rank tests were applied to assess differences in phenologic metrics for crops growing on low and high PAWC soils. This allowed us to rank phenologic metrics for use as indicators of soil PAWC. The results show a significantly higher GreenUpSlope (p<0.00008) and maximum NDVI (p<0.0004) in low PAWC soils. This indicates that crop phenology derived from MODIS satellite imagery may provide useful information about soil water conditions, which would allow improvements to the spatial detail in soil maps. (C) 2015 Elsevier Ltd. All rights reserved.