Drought Management Guidelines

The objective is to provide the methodology to be applied for the characterization of meteorological and hydrological droughts. The correct drought characterization provides decision makers with a measurement of abnormal weather variability, so that protection from possible impacts may be implemented.

Drought is a three-dimensional phenomenon that can be characterized by its severity or intensity, duration, and geographic extent. Drought characterization is complex and there are a wide range of meteorological or hydrological indices or indicators that can be used. It requires an accurate selection of drought identification methods and/or of drought indices, able to describe in a synthetic and clear manner the evolution of drought conditions in space and time. Each one has its own merit and they are often supportive of each other. A combination of indices and indicators is usually the preferred option.

Drought indices can be used to describe all types of droughts (that is meteorological drought: deviation from the normal meteorological conditions, hydrological drought: deviation from the normal hydrological conditions, agricultural droughts: deviation from the normal soil moisture conditions for crop growth and socioeconomic drought: deviation from the normal level of availability of water for fulfilling societal needs.

The indices for drought characterization have to comply the following requirements: (a) can be calculated from data available from actual data collection systems; (b) have a priori and direct relation with vulnerable social, economic and environmental systems; and (c) can be used for predictions and early monitoring systems.

The expected outcome is the characterization of the meteorological and hydrological drought periods in the historical record in each geographical unit.

Drought management depends on indices to detect drought conditions, and thresholds to activate drought responses. Indices and thresholds are important to detect the onset of drought conditions, to monitor and measure drought events, and to quantify the hazard.

The appropriate drought index is selected according to the type of drought. Indices may be considered as general or specific depending on the utility for which they have been devised. It is understood that this distinction is difficult. Some of the indices, however, are more appropriate for monitoring and some for the analysis of historical drought events.

Drought indices are essential elements for drought monitoring since they summarize the complex interaction between climatic variables and related processes (e.g. soil water moisture). Use of indices allow a quantitative assessment to be made of the climatic anomalies in terms of intensity, spatial extent and frequency, and favour the exchange of information about drought conditions among decision makers as well as the public.

The availability of a large number of indices is mainly due to the difficulty in defining unequivocally a drought phenomenon. The current common orientation consists of the application of a group of different indices within a monitoring system of hydro-meteorological variables and water resources availability provided by public "monitoring centres". The main purpose of "monitoring" centres is to support decision makers in timely recognizing drought onset.

Different indices and methods have been proposed since the '60s to identify and monitor drought events. Some of the indices refer to meteorological drought and are based on precipitation series, while others are oriented to describe hydrological or agricultural drought or water shortages in urban water supply systems. Table 4 presents a summary of some of the main indices that can be applied to drought characterization and monitoring.

Drought Indices	Data needed	Category of use
Deciles	Precipitation	Meteorological
Standardized Precipitation Index (SPI)	Precipitation	Meteorological, used for monitoring and forecasting
Rainfall Anomaly Index	Precipitation	Meteorological, sensitive to extreme events
Reconnaissance Drought Index (RDI)	Precipitation, Potential Evapotranspiration	Meteorological
Run Analysis	Precipitation, streamflows	Meteorological and hydrological, for spatio-temporal analysis of historical events
Palmer Drought Severity Index (PDSI)	Precipitation, Temperature, Soil Moisture (Available Water Content)	Meteorological, effective in agriculture, used in historical analysis and risk analysis
Palmer Hydrological Drought Severity Index (PHDI)	Precipitation, Temperature, Soil Moisture Conditions	Hydrological, effective in monitoring
Palmer Moisture Anomaly Index(Z-Index)	Precipitation, Temperature, Soil Moisture Conditions	Agricultural
Surface Water Supply Index (SWSI)	Snowfall, Precipitation, Streamflow, Reservoirs	Hydrological, effective when snow is important
Crop Moisture Index (CMI)	Precipitation, Temperature, Soil Moisture Conditions	Agricultural
Soil Moisture Anomaly Index (SMAI)	Soil Moisture Conditions, Potential Evapotranspiration, Potential Runoff	Hydro-Agricultural
Normalized Difference Vegetation Index (NDVI)	Satellite images	Natural resources, agricultural

The most commonly applied drought indices include the Standardized Precipitation Index (SPI), the Palmer Drought Severity Index (PDSI) and Deciles due to their simplicity. It was concluded that the easiest index to use for monitoring purposes is the SPI, which is based on a single meteorological parameter (precipitation) and the RDI that also includes evapotranspiration. Recent advances in remote sensing provide products that have a large potential as drought indices. The NDVI is widely used for monitoring and forecasting crop production world-wide and by agricultural insurance companies.

Because hydrometeorological parameters are measured at certain stations and decisions should be taken in most cases at basin level, spatial integration is required in the case of applications of the methodology for water and agricultural management decisions. Spatial integration at a level of a small basin or sub-basin may be implemented by calculating the weighted mean of the parameters involved. Weight in this approach is the area represented by each station. The spatial extent of drought is estimated based on comparisons of the affected area with a threshold referred to as "critical area". A promising method with the flexibility to use various area thresholds associated with each severity level of drought is based on plotting the percentage affected area against each level of severity of drought.

As far as the time step is concerned characterization of drought can be based on an annual time step accompanied by other shorter time steps (e.g. six months, three months) or any other time duration tailored for the specific application. The selection of the time step applied is a crucial element in the analysis as well as the selection of the threshold for each index.

Box 4: Examples of spatial application of the SPI drought index in Sicily, Italy, during a drought year (2002) and a normal year (2005). Maps correspond to the situation in January of each year.

Box 4- Figure

The objective of the assessment of drought characteristics is to evaluate the severity and duration of droughts that can occur in a given area or region in probabilistic terms. Such an assessment is useful for the analysis of past droughts, and to define "design droughts" of a fixed return period, that assist in the analysis of risk.

Box 5: Statistical properties of drought

Among the different proposed methods for characterizing droughts, the run method has found widespread use, due to the objectivity in the definition of drought. Furthermore, the method allows an analytical derivation of the probability distributions of drought characteristics to be carried out, thus overcoming the limits of an inferential approach due to limited sample lengths. The run method can be applied to a hydrological series of interest, either at yearly or sub-yearly time scales (e.g. precipitation, streamflows, etc.) assuming as a threshold a value representative of the demand level. The method can also be extended to the case of regional droughts, by introducing a threshold representative of the areal extension of deficits.

Box 5- Figure

Drought identification using the theory of runs (drought periods in red colour)

- Drought indices are not a goal, but a means to identify and analyse droughts.
- Even though computation of indices can be complex, the resulting outputs should be presented in a simple format.
- Some of the indices include relevant meteorological and hydrological information, but do not consider water uses in the basin.
- A clear criterion to identify droughts is not universal. All indices are sector/system-specific, therefore multiple indices should be used to characterize drought.
- Recent remote sensing-based indices may have a large potential as drought indices, especially where other sources of data are limited.
- Meteorological drought indices may not correlate well with historical drought impacts, due to the effect of storage in regulated systems (e.g., over year storage). On the other hand, drought indices are very useful in rainfed conditions to forecast agricultural production.
- All indices are sector/system specific.
- The optimal approach for using indices is to calibrate them with observed impacts, risk level, and vulnerability.