Rapidly assessing the 1997 drought in Papua New Guinea using composite AVHRR imagery

During 1997 Papua New Guinea (PNG) experienced an intense drought. Emergency famine relief operations provided many subsistence agricultural communities with food, water and health provisions during the height of the drought. The locations of relief operations were based on a rapid and spatially explicit extensive field survey conducted at the height of the drought for all PNG. We have tested the utility of composite Advanced Very High Resolution Radiometer (AVHRR) data to assist in a rapid assessment of drought conditions in PNG. Composited data were used to provide a means to overcome the frequent cloudy conditions that exist in PNG. To assess the drought we divided land surface temperatures(T-s) by the Normalized Difference Vegetation Index(NDVI). The ratio (T-s/NDVI) increases during times of drought. This is due to the increase in T-s associated with more net radiation being partitioned into the sensible heat Aux and the decrease in NDVI associated with decreasing amounts of plant cover. A time series of T-s/NDVI is a rapid indicator of the drought at the country and province level. We calculated the integral under the T-s/NDVI curve for 1997, denoted integral (Dec)(Jan) T-s/NDVI, which allowed us to validate the assessment in two ways. Firstly, it was compared with field-based assessments of food supply (undertaken at the height of the drought) which classified areas into five grades. Plotting integral (Dec)(Jan) T-s/NDVI against the percentage area of each of the 14 mainland provinces experiencing food supply problems (so lives were at risk) showed a strong positively correlated relationship (r(2)=0.818). Secondly, for seven meteorological stations plotting integral (Dec)(Jan) T-s/NDVI against cumulative 1997 rainfall showed an inverse relationship (r(2)=0.809). Both forms of validation show that the composite AVHRR T-s/NDVI ratio provides a rapid means to assess drought conditions in a cloudy environment such as PNG.