Improving Princeton Forcing Dataset over Iran Using the Delta-Ratio Method

In this study, we corrected the bias in the Princeton forcing dataset, i.e., precipitation, maximum and minimum temperatures, and wind speed, by adjusting its long-term mean monthly climatology to match observations for the period 1988–2012 using the delta-ratio method. To this end, we collected meteorological data from 97 stations covering the domain of Iran. We divided Iran into three climatic zones based on the De Martonne classification, i.e., Arid, Humid, and Per-Humid zones, and then applied the delta-ratio method for each climatic zone separately to adjust the bias. After adjustment, the new datasets were compared to the observations in 1958–1987. Results based on four skill scores, including the Nash Sutcliffe efficiency (NSE), percent bias (PBIAS), root-mean-square error (RMSE), and R², indicate that the adjustment greatly improved the quality of the gridded dataset, specifically, precipitation, maximum temperature, and wind speed. For example, NSE for annual precipitation during the validation time period increased from −0.03 to 0.72, PBIAS reduced from 29.2% to 6.6%, RMSE decreased by 182.44 mm, and R² increased from 0.06 to 0.75. Assessing the results in different climatic zones of Iran reveals that precipitation improved more significantly in the Per-Humid zone followed by the Humid zone, while maximum temperature improved better in the Arid areas. For wind speed, the values improved comparably in the three climate zones. However, the delta values for monthly minimum temperature calculated during the adjustment time period cannot be applied in the validation time period, due to the fact that the Princeton climate data cannot follow the behavior of minimum temperature during the validation phase. In short, we showed that a simple bias adjustment approach, along with minimum observed station data, can significantly improve the performance of global gridded datasets.