Abstract/Summary

This research deals with the assessment of drought over Iraq, a country which has been affected by recurrent droughts of medium to long duration, with serious socio-economic consequences. The overarching aim of this work is to identify reliable large-scale drought monitoring and assessment methods, using a range of freely available meteorological and remote sensing data, as well as model simulations of the water balance. The thesis starts with an overview of Iraq’s climate, soil, land use and socioeconomic issues, as well as an inventory of commonly used methods to assess and express drought. Historical droughts in Iraq have been studied between 2001 and 2013, using a combination of meteorological drought indices, remote sensing products, and water balance estimates by the SWAP model and ERA-Interim reanalysis. Drought is assessed for a number of key land surface types (desert, rangeland, agricultural land, and marshland), with the first three representative of the different climatic zones in Iraq. Their evolutions of drought have been compared and contrasted. The research uses Standardised Precipitation Index (SPI) and Standardised Precipitation Evaporation Index (SPEI) drought indices, derived from ERA-Interim/in-situ data of rainfall and temperature, that are applied to evaluate meteorological droughts in Iraq. The effect of the meteorological droughts has also been analysed using land surface temperature (LST), Normalised Difference Vegetation Index (NDVI) and near-surface soil moisture content (SMC), derived from remote sensing data, in isolation and together with SPI/SPEI, for the years 2001 to 2015. NDVI has been used widely to detect changes in vegetation extent; LST was employed as a proxy of land surface evapotranspiration. NDVI was obtained from MOD13A2 products (16-Day L3 Global 1km SIN Grid VI datasets), which were designed for vegetation. LST was obtained through MOD11A2 products available at a spatial resolution of 1km and a temporal resolution of 8 days. SMC was derived from the Soil Moisture and Ocean Salinity (SMOS) product at a resolution of 40 × 40 km. The evolution NDVI and SMC overall followed that of the drought indices, but the interannual variations of seasonal LST courses were not deemed useful for drought assessment. Furthermore, regression analyses were conducted between SPI3/SPEI3 and NDVI as well as LST to investigate the potential of drought inference from the sign and strength of the correlation coefficient of the slope. This approach appeared promising for the marshlands, but less so for desert, rangeland, and agricultural regions. Finally, the soil water balance Assessment (SWAP) model was used, driven by in-situ and ERA-Interim data. Water balance components for each land surface type were studied over time, to determine the effect of meteorological droughts on the variation of predicted ecosystem and agricultural system’s hydrological behaviour. Comparisons were also made between the various land surface types, and between SWAP and ERA-Interim actual evapotranspiration estimates. The outcomes showed that the lowest mean actual evapotranspiration and water storage was found between 2008-2010, and 2012, as a result of lack in rainfall, whereas the monthly averages of actual evapotranspiration and water storage were the highest in 2013 and 2014. The data of the surface latent heat flux (ERA-Interim), actual evapotranspiration (SWAP), and soil moisture corresponded well for the desert and rangelands but not for the agricultural region. This is the result of the fact that ERA-Interim does not cater for (irrigated) crops whereas SWAP does. The thesis concludes with recommendations with regards to the usefulness of the various (meteorological and remote sensing-based) drought indices, and combinations thereof, for assessment of drought in Iraq and similar climatological conditions.