Application of MODIS and Google earth data to assess the health of the Mitchell grasslands through web based GIS

In North Western Queensland and the Central Eastern portion of the Northern Territory lie the Mitchell Grasslands. Covering approximately 320 000 square kilometres, the area is grazed by approximately 12 million cattle. The use of the Grasslands for grazing pasture has placed pressure on the natural processes of the area. Also, working in the unpredictable conditions that occur from year to year has certain challenges for graziers and land managers alike. By aiming to find a spatial solution is assess the health of the grassland, and applying the knowledge gained, better economic, social and environmental outcomes can be achieved. This study has used the remote sensing method of multi-spectral imaging to generate ‘greenness’ or vegetation index (VI) plots of a selected portion of Mitchell Grassland. By comparing the red and near-infrared bands detected by the Moderate Resolution Imaging Spectroradiometer (MODIS) satellites, Terra and Aqua, differentiation can be made between sections of healthy and stressed vegetation. Images were collected over the 17 year period that MODIS has been operating, at an interval of 3 months. From these images GIS analysis and data manipulation was completed and a number of statistics were calculated including average normalised difference vegetation index (NDVI) value and percentage of pixels above a specified NDVI value. The Queensland Department of Science, Information Technology and Innovation (DSITI) along with its regional climate/meteorological program, The Long Paddock run GRASP (Grass Production) model, as well as the spatial implementation of the results called Aussie Grass. Data produced includes pasture biomass, relative pasture biomass, pasture growth, rolling average relative pasture growth 1, 3, 6, 12 and 24 months, pasture growth seasonal probability, pasture curing index, grass fire index, rolling average relative rainfall 1, 3, 6, 12 and 24 months and total monthly rainfall. Pasture coverage data was extracted and compared with the percentage of NDVI values above a specified range to estimate a comparison. Some positive results were seen from this analysis with regression analysis showing a weak correlation. The NDVI data, particularly the average over a seasonal and annual basis were also compared to weather and climate data including rainfall averaged over 3 nearby Bureau of Meteorology stations and SOI data for the same time period. A strong correlation was seen with the rainfall data, confirming the findings of previous work, while a weak correlation was found between the NDVI and the SOI. This is likely due to the small study area in terms of the area that SOI covers. Finally, each NDVI snapshot of the study area was processed into a colour-scaled image and presented in the form of a KML time-series for easy use, sharing and editing. It is hoped that this study will inspire further investigation into areas such as the analysis of bushfires using this technology, comparison with other datasets including the Aussie Grass pasture growth rates and the use of LIDAR to analyse grassland dynamics as well as grass height/availability of feed