Application of Drought Indices for the Changed Climate

Drought conditions are commonly assessed using the drought indices. Of these, the Standardized Precipitation Index (SPI) and the Palmer Drought Severity Index (PDSI) are the most frequent. In addition to SPI (1 to 24 months aggregation) and PDSI, the Z index is also used in the Czech Republic. It is closely related to PDSI but does not account for the persistence; rather, it characterizes the immediate (for a given week or month) conditions. While the SPI is based solely on precipitation data, the PDSI and Z indices are based on precipitation and temperature data and on the available water content of the soil. For climate change impact studies, we consider PDSI and Z to be better indicators as they account for the changes in both temperature and precipitation. The three indices are calculated by programs available from the National Drought Mitigation Center and the Department of Computer Science and Engineering, both located at the University of Nebraska-Lincoln. In the original versions of the three indices, their values are calibrated using a single input series data so that the range of the values is approximately the same for all input weather series (self-calibrated indices). To allow assessment of the climate change impacts, a modification was made: the indices are now calibrated using the present climate (learning) weather series and then applied to the future climate weather series (relative indices). In assessing impacts of climate change, we will analyze not only the shifts in the values of the indices but also changes in the drought event characteristics (e.g. frequency, mean duration, accumulated intensity, peak intensity). The drought events are defined here as continuous periods in which the index: (a) does not exceed the selected upper threshold value, and (b) falls at least once below the lower threshold value. Threshold values of 0 and -1 are used to define the SPI based drought spells, -1 and -3 are used to define PDSI-based and Z-based drought spells. Two approaches will be used to assess the impacts: (1) The drought indices will be derived from European and U.S station weather series representing the present and future climates. In both cases, the present-climate series will serve as a learning series. Present climate will be represented by the observed weather series, the changed-climate weather series will be obtained by applying GCM-based change factors to the observed series. This approach does not allow for the possible changes in inter-monthly variability which could affect drought spells, therefore only shifts in the changes in indices values will be analyzed. (2) The drought indices will be derived from the grid-related GCM-simulated (whole globe) surface weather series for 1961-2000 (present climate) and 2060-2099 (changed climate). This will allow to analyze not only shifts in the values of the three indices, but also changes in drought event characteristics due to changes in inter-monthly climate variability (assuming that it is reproduced by the GCMs). In both approaches, multiple GCM simulations (run at SRES-A2 and SRES-B2 emission scenarios) available from the IPPC data distribution center will be used to account for the inter-GCMs uncertainties. Acknowledgements: NATO Science Program - project EST-CLG 979505 (Drought as the limiting factor of the cereal production).