The first challenge towards an effective drought monitoring system in Latin America is to build awareness of drought as a recurrent phenomenon. It is known that some regions have greater exposure to drought than others and we do not have the capacity to alter such exposure.
However, we can collect and analyze multiple data, such as rainfall, soil moisture, vegetation stress, groundwater levels or socio-economic data on a variety of time and geographical scales, and provide the information necessary for identifying droughts and estimating their frequencies, as well as to formulate actions to mitigate their impacts on human activities.
As a temporal and recurring phenomenon, a drought should always be defined as an abnormal water supply condition relative to some long-term-average state, e.g. current monthly precipitation, weekly balance between precipitation and evapotranspiration, etc. Because deficiency of rainfall is always the driver of a drought, practically all drought indices use precipitation either singly or in combination with other meteorological variables, depending on the type of requirements. For example, agricultural drought starts only when the duration and intensity of meteorological drought increases and disturbs the expected development of vegetation, which can be measured by soil moisture conditions or vegetation greenness state from remote sensing images.
Drought indices computed on a monthly or weekly basis seem to be the most appropriate for monitoring the effects of a drought in situations related to agriculture, fresh water supply and ground water abstractions. A combined time series of different drought indices provides a framework for evaluating the drought parameters of interest.
Within the context of the EUROCLIMA programme of the European Union and Latin America, the JRC DESERT Action is responsible for developing methodologies and tools for monitoring and assessing drought events in their distinct typologies and for mapping characteristic drought frequencies for the region.
To undertake this task, the Standardized Precipitation Index (SPI), a Standardized Vegetation Index (SVI) and a Drought Frequency Index (DFI) will be computed at the continental level on monthly, ten-daily and yearly basis, respectively, and will be made available to the LA community using a web-map server, i.e. the EUROCLIMA DLDD Information System.
The SPI was designed to be a relatively simple, spatially invariant and probabilistic year-round index applicable to water supply conditions. The SPI is based on precipitation alone and is defined as the number of standard deviations that the observed accumulated rainfall at a given location and timescale deviates from the long-term normal conditions.
Positive SPI values indicate greater than the median precipitation, and negative values indicate less than median precipitation. The fundamental strength of the SPI is that it can be calculated for a variety of timescales, enabling water supply anomalies relevant to a range of end users to be readily identified and monitored. For example, SPI to monitor short-term water supplies, such as soil moisture, is important for agricultural production, and SPI for longer timescales is important for, amongst others, groundwater supplies and reservoir levels.
At the continental level, the SPI is computed with the GPCC (Global Precipitation Climatology Centre) monthly gridded precipitation from the DWD (Deutscher Wetterdienst) at 1° spatial resolution.
SVI is designed to show the effects of drought on vegetation greenness over short-time periods, usually a week or 10 days. Intuitively, the SVI is an estimate of the “probability of occurrence” of vegetation greenness. SVI is based on Vegetation Indices (VI) alone, such as the Fraction of Absorbed Photosynthetically Active Radiation (fAPAR), Normalized Difference Water Index (NDWI) or Normalized Difference Vegetation Index (NDVI), and computed as the z-score for each VI image pixel location.
The z-score is a deviation of the current vegetation greenness from the long-term mean in units of standard deviation, calculated from the VI values for each pixel location for each short-time period. Low SVI values indicate poor vegetation condition that could be the result of climate conditions; high SVI values might reflect ideal climate growing conditions so that vegetation greenness is higher than encountered in other years. At the continental level, the SVI is computed from 10-daily NDWI and NDVI, both computed from the SPOT-VEGETATION satellite data, and from the fAPAR, derived from MERIS satellite data at 1km spatial resolution.
The DFI (see map) is a new index, still in development, proposed by the JRC DESERT Action to estimate the long-range likelihood trend of drought occurrences and their spatial and temporal persistency. The DFI should give important structural information on drought that can be used as a basis for regional long-term risk assessment and prediction. In short, historical low rainfall regimes are computed for each location based on the empirical cumulative distribution of monthly precipitation totals and the nonparametric Fisher-Jenks optimal classification algorithm.
The time-series of water supply deficiencies is then used to compute Markov Chains of historical monthly drought events and the marginal distributions of the respective probability transition matrices are used to derive long-term drought frequencies. At the continental level, the DFI is computed with the GPCC (Global Precipitation Climatology Centre) monthly gridded precipitation from the DWD (Deutscher Wetterdienst) at 1° spatial resolution.