Abstract:

Pakistan is in South Asia (60°-80°N, 23°-37°E) with a total area of 88×10⁴ km². The southeastern part of Pakistan is mountainous and hilly with a complex topography and different climate types. In summer, this area is controlled by subtropical high pressure, which results in hot and dry weather with a small amount of precipitation. The average annual rainfall is less than 250 mm, and a quarter of the area receives less than 120 mm. Because of the high summer temperatures, Pakistan has been greatly affected by long-term drought, which has severely restricted its social and economic development. In this study, we used the Global Inventory Modeling and Mapping Studies Normalized Difference Vegetation Index 3^rd Generation (GIMMS NDVI3g) dataset from 1982 to 2013 to construct the vegetation condition index (VCI). We then used the VCI to analyze the range and frequency of drought in Pakistan for the past 32 years. The drought frequency was used to characterize the risk of drought based on sources from the Pakistan Statistics Bureau and reports in the literature. Information on livestock, soil, crops, and irrigation were used to evaluate the drought vulnerability, exposure, and disaster prevention and mitigation capabilities of Pakistan. Additionally, ERA-Interim data that had been verified by 10 meteorological stations in Pakistan were used for temperature and precipitation reanalysis. The influencing factors of the drought range were determined using partial correlation coefficients, and the partial least squares method was used to identify the dominant influencing factors for the spatial patterns of regional drought risk. The following results were obtained. (1) During the study period, the multiyear average drought range differed for each month. The drought range was larger from January to September than from September to December (area ratio was greater than 40%). The drought range was smaller from May to August (area ratio was less than 30%). The increase in temperature and decrease in precipitation from May to June caused the arid area to increase. (2) In the study area, the drought frequency was low from May to August. The frequency was higher in September for the woodlands and grasslands of the western mountainous areas. The frequency was higher in January and September than in December for the eastern plains. (3) The spatial pattern of drought risk was mainly determined by the drought frequency, crop yield, proportion of large livestock, and water storage capacity of the soil. Among these, the drought frequency had the highest impact on the regional drought risk. The livestock ratio and the water storage capacity of the soil indicated a higher drought risk in the transition zone from mountains to plains in the northern area and the southeastern area covered by natural vegetation. Irrigation reduced the drought risk in the valley area with flat terrain. Additionally, the natural vegetation in the northern mountainous area reduced the drought risk as well. In this study, we analyzed the long-term drought characteristics in Pakistan and evaluated the drought risk due to natural and human factors. This research can provide useful information for drought and disaster risk management in Pakistan.

Key words: drought, risk assessment, normalized difference vegetation index(NDVI), vegetation condition index(VCI), Pakistan