Feasibility of Drought Prediction
Weather patterns in many parts of the world appear to be related to different phases of the El Niño/Southern Oscillation (ENSO; chapter 3) cycle. The existence of such linkages is now being used in operational early warning systems, such as FEWS, to forecast rainfall patterns for the coming cropping season. The basis for forecasting is that a particularly strong linkage exists between the warm-ocean phase of the ENSO and drought in southern Africa. Such correlations have proved useful in tropical areas, and it may soon be possible to predict, for Southern and Eastern Africa, certain climatic conditions associated with ENSO events more than a year in advance.
In the North Africa and West Asia, no significant relationships have been confirmed between droughts and ENSO events. This is probably because the effect of this global ocean-weather linkage is substantially modified by more localized weather phenomena, such as the NAO, but also by very site-specific factors, in particular the topography and the nearness of large desert landmasses in the Sahara and the Arabian Peninsula. Nevertheless, according to Ward et al. (1999), sufficient evidence exists to suggest an
ENSO influence on sea-surface temperatures in the northern Atlantic and on late-season (March-April) precipitation levels in northwestern Africa. Particularly in Morocco, west and north of the Atlas Mountains, late-season precipitation tends to be below normal during warm ENSO years.
A major research need is to investigate whether the NAO, like ENSO, is a coupled ocean-atmosphere phenomenon as opposed to a random atmospheric phenomenon. If the NAO were a coupled phenomenon, it would enhance the use of an NAO index in drought early warning systems (Cullen and deMenocal, 2000). However, as Iglesias (2001) points out, the large variability of climate in the region, spanning time scales from the intrasea-sonal to the decadal, poses particular challenges for the management of agriculture. It is therefore expected that, even if this research leads to a successful outcome, the contribution of seasonal forecasts to the stabilization of agricultural production in the Near East will be relatively modest as compared to other management strategies adapted to highly variable climates.
Spatialization of Drought in Data-Insufficient Areas
The Near East is a region with relatively sparse and heterogeneous climatic data coverage. With the exception of Turkey, which has a good and homogenous coverage throughout the country, most climatic stations in the region are concentrated in coastal and agriculturally important areas. Rangeland (arid) areas and deserts are very poorly covered. Due to the pressure of the population increase, these areas are becoming increasingly important from an ecosystem function perspective. As the region is also diverse in terms of landscapes and topography, temperature regimes are not uniform, which needs to be taken into consideration while using drought indicators based on the water balance.
In such environments characterized by high spatial variations in moisture and temperature regimes, the delineation of drought can be considerably improved by advanced methods of spatial interpolation. Several statistical techniques are now available that make use of digital elevation models (DEM) to improve the spatialization of climatic parameters (De Pauw et al., 2000). In view of the strong linkages between climatic variables (especially temperature, but also rainfall, humidity, and sunshine) and topography, the most promising techniques for spatialization in climatology are multi-variate approaches because the latter permit the use of terrain variables as auxiliary variables in the interpolation process. In contrast to the climatic target variables themselves, which are only known for a limited number of sample points, terrain variables have the advantage that they can be known for all locations in between, which increases the precision of the interpolated climatic variables significantly. Co-kriging (e.g., Bogaert et al., 1995) and co-splining (e.g., Hutchinson, 1995) are methods that in most cases lead to excellent interpolations. ICARDA has successfully combined the co-splining approach of Hutchinson with the GTOPO30 digital elevation model (Gesch and Larson, 1996) for regional-level mapping of various basic and derived climatic variables at 1-km resolution (De Pauw, 2002). Another important tool for spatialization is remote sensing. Remote sensing has become a standard tool in most food security early warning systems, such as FEWS, GIEWS, and MARS (Monitoring Agriculture with Remote Sensing, a project of the Joint Research Center of the European Commission; http://mars.aris.sai.jrc.it). This development has been promoted by the decreasing costs of satellite data products and image analysis tools, the difficulty of obtaining timely climatic data, and significant correlations between soil moisture status or biomass productivity and some parameters derived from spectral analysis (e.g., normalized difference vegetation index). Further details on remote-sensing techniques for drought monitoring are provided in chapters 5-8.
A first task for an early warning system is to understand the spatial variations of drought risk. If good time series data exist for spatially well-distributed climatic stations, drought risk can be spatialized from the probability surfaces of selected drought indicators. However, for drought planning it is also essential to go beyond the symptoms of drought, as they appear from the meteorological or hydrological records. As experienced in the region, access to and the stability of the natural resource capital (particularly natural vegetation, climate, soil, and irrigation water) are major determinants of the resilience of rural livelihood systems against drought. Understanding the underlying causes of vulnerability and anticipating the impact of drought thus requires an integrated approach, which considers both the differences in agroecological and socioeconomic characteristics between different areas.
The basis for mapping vulnerability to drought could be a spatial framework of combined agroecological zones and production systems zones. The agroecological zones can be established by integrating available climatic, soil, terrain, and land cover digital data sets. The production system zones can be derived from remote sensing in combination with farming systems information. By integrating the spatial agroecological and socioeconomic data, "agroecozones" can be established, which have unique characteristics in terms of climate, soil, and water resources, population characteristics, and livelihood systems. The agroecozones offer a useful framework for selecting sampling areas as part of a regular drought-monitoring program.
Was this article helpful?