A method for computing the monthly Palmer Drought Index on a weekly basis: comparing data estimation techniques

In 1965, Wayne Palmer presented an index that he hoped would be universally applicable for monitoring drought and wet spells in the United States heartland. It was hailed as a major improvement in drought monitoring up to that point and saw widespread adoption. Palmer developed his index using monthly data, and the National Oceanic and Atmospheric Administration's (NOAA) National Climatic Data Center (NCDC) computed historical Palmer Drought Index (PDI) values on a monthly timescale back to 1895 for 344 climate divisions in the contiguous United States. In order to create a timelier PDI for operational purposes, NOAA's Climate Prediction Center (CPC) computes the PDI on a weekly basis for these same climate divisions. The weekly PDI data base starts only some three decades ago, so the weekly values are utilized in conjunction with the monthly values to put current climate anomalies into a century-scale historical perspective. Because of time step and other methodological differences, the values for the monthly and weekly PDIs for the same ending dates sometimes disagree significantly. The research presented in this paper is an outgrowth of an effort to resolve the differences between the weekly and monthly PDIs operationally created by NOAA. A methodology is presented for estimating mean monthly temperature and total monthly precipitation before the month is over so that a monthly PDI can be computed from this data on a weekly basis. The approach involves combining observed temperature and precipitation data for the month to date with daily normals for the remainder of the month. It is demonstrated to be superior to two other techniques, a persistence method and a moving window method. This normals technique allows the computation of an operational monthly PDI on a weekly (or even daily) basis which is consistent with historical drought indices of the recent past.