Agricultural Drought Indices - US Department of Agriculture

hydrological characteristics, and crop phase. Another limitation of more complex agricultural
drought indices is the great complexity of agriculture itself, with several crops being cultivated
during the growing season in the same region and having different stages of development at a
given time. It is very well known that the same water deficit will impact yield differently if it occurs in
distinct phases of crop development.
Another source of uncertainty for agricultural drought indices that are based on WB outputs is
related to the different ways to estimate ETP. Different ETP methods will result in different values
for the same weather conditions. This will make agricultural drought indices vulnerable to these
methods, in terms of the right dimension of the drought index. This is a problem when the Penman-
Monteith method cannot be applied because of lack of data. So, agricultural drought indices based
on ETP or ETa estimated by different methods are not comparable. The ideal would be the
Penman-Monteith FAO56 model (Allen et al. 1998); however, this is not always possible, since this
method requires a complete meteorological dataset, including net radiation. When alternative ET
estimate models are employed, one should pay attention to the characteristics of the models.
Some of them, like Thornthwaite, tend to underestimate ETP during dry periods, whereas the
Hargreaves and Samani method can overestimate ETP during the wet season. The type of soil,
the depth of the roots, and the resulting SWHC is another source of uncertainty for the WB-based
agricultural drought indices. The greater the SWHC, the smaller the impact of a given drought on
crop yield.
Even considering their limitations, the WB-based agricultural drought indices are the best option for
monitoring droughts for agriculture, presenting better correlations to yield losses than indices
based only on rainfall data.
Conclusions
In Brazil, several national and regional M&HSs have different ways to monitor drought under
agricultural perspectives. These M&HSs estimate water balance by four different methods:
Thornthwaite and Mather, MUSAG, hydrological with satellite data, and hydrological with observed
data. The method of Thornthwaite and Mather (1955) is the most used by the National Met Service
(INMET and AGRITEMPO) and other regional agrometeorological services. In the state of São
Paulo, several agricultural drought indices are in use, with the majority of them based on
Thornthwaite ETP and Thornthwaite and Mather’s WB. Even considering the strengths of the WBbased
agricultural drought indices, the WB, determined by any method, will depend on some
critical factors, such as the ETP method, the SWHC adopted, crop/variety type and phase, and
crop management. These factors will lead to different agricultural drought index values, which will
require calibration and testing of them for each location and crop condition. However, agricultural
drought indices based on WB outputs are expected to have better relationships with crop yield
losses than meteorological drought indices, which are based only on rainfall.
References
Allen, R.G., L.S. Pereira, D. Raes, and M. Smith. 1998. Crop Evapotranspiration.Guidelines for
Computing Crop Water Requirements. FAO Irrigation and Drainage Paper, No. 56. FAO,
Rome, Italy.
Berlato, M.A. and A.P.A. Cordeiro. 2005. Variabilidade climática e agricultura do Rio Grande do
Sul. Pages 43-58 in As estiagens e as perdas na agricultura – fenômeno natural ou
imprevidência? (Nascimento, A.M.N., Silveira Filho, I.L., Berton, A.L. e outros, eds.).
Federacite, Porto Alegre, Brasil.
Blain, G.C. 2005. Avaliação e adaptação do índice de severidade de seca de Palmer (PDSI) e do
índice padronizado de precipitação (SPI) às condições do estado de São Paulo. Instituto
Agronômico, Campinas, Brasil.
Boken, V.K. 2005. Agricultural drought and its monitoring and prediction: Some concepts. Pages 3-
10 in Monitoring and Predicting Agricultural Drought (V.K. Boken, A.P. Cracknell, and R.L.
Heathcote, eds.). Oxford University Press, New York.
70