Comparative Evaluation of TRMM Estimated Rainfall Amounts and Rainfall Recorded By Ground Stations in Lake Urmia Basin

Rainfall is one of the most important elements in determining the climate that has been regarded by experts in various fields. This element with spatial and temporal changes is one of the most important inputs of the hydrological systems that is necessary its study and measurement in several different conditions, such as Climate modeling, climate change, prediction of atmospheric condition, study of runoff, groundwater, flood modeling. Therefore, it is important to accurately estimate of its value.
A comprehensive estimation of precipitation remains one of the most difficult observational challenges of meteorology, especially during convective rainfall events, since these types of precipitation events develop quickly and do not always last very long. Although rain gauges provide a direct measurement of rainfall, rain gauge networks will always be too coarse. Rain gauges are also unevenly distributed and, most importantly, they provide point source data and not a representation of a spatial domain.
Radar rainfall can be used to provide an indirect measurement of rainfall, but then the radar systems need to cover the entire area of interest, be well correlated and have appropriate radar rainfall relationships according to the type of precipitation. For most developing countries and even more so for least developed countries radar systems will remain too expensive and difficult to maintain and thus not a feasible option for this purpose.
Satellite based estimates of rainfall are not as accurate as gauges or radar rainfall, but has the advantage of high temporal resolution and spatial coverage, even over oceans, in mountainous regions and sparsely populated areas. In areas where there are very few rain gauges and no radar systems, satellite-derived rainfall can be a critical tool for identifying hazards from smaller-scale rainfall and flood events. Satellite based Precipitation Estimators should not be considered as a replacement for radar rainfall estimates and gauges, but as a complement to these fields, if at all possible.
The purpose of this study was to assess the comparative evaluation of TRMM estimated rainfall amounts and rainfall recorded by ground stations in Lake Urmia Basin.
The Lake Urmia Basin is located between 35°40ʹ to 38°30ʹ latitude and 44°14ʹ to 47°53ʹ longitude in northwest Iran and covers an area of 51,800 km2which composes 3.15 % of the entire country and includes 7 % of the total surface water in Iran. The Lake Urmia is the largest lake in the country and is also the second hyper saline lake (before September 2010) in the world and it is an important natural asset with considerable cultural, economic, aesthetic, recreational, scientific, conservation and ecological value. The lake basin includes 14 main sub basins that surround the lake with the areas varying from 431 to 11,759 km2.
In this study, the daily and hourly rainfall data of 16 synoptic stations in Lake Urmia Basin during the period 2005 to 2011 and the three-hourly rainfall rate of TRMM 3B42-V6 at 0.25 ° resolution are used. The Run Test was used to investigate the homogeneity of data. To study the performance of TRMM in rainfall estimation, the performance measures were used such as Proportion Correct (PC), bias, Probability of Detection (POD), False alarm Ratio (FAR), and Heidke's Skill Score (HSS) that are defined by using the standard 2 × 2 contingency tables. To validation of TRMM rainfall data, the statistical criteria like the mean error (ME), mean absolute error (MAE) and root mean square error (RMSE) were used and calibration of TRMM rainfall amounts were used by ground station's data and linear regression.
Discussion and TRMM rainfall amounts downloaded by netcdf format and were converted to the raster map in ArcGIS 10. To study the performance of TRMM in rainfall estimation, the performance indices of bias, Probability of Detection (POD), False alarm Ratio (FAR), Percentage of Corrects (PC) and Heidke's Skill Score (HSS) were used. Average of these indices in the basin was respectively 0.60, 0.52, 0.13, 0.68 and 0.39. As it can be seen, Bias value is less than one. So we can conclude that TRMM underestimated rainfall. High values of POD and low values of FAR indicated that this satellite has acceptable performance in rainfall estimation. According to PC, TRMM Has estimated Properly In 60% to 70%. According to HSS amounts, it can be said that there is relative match between satellite data and observed data. To validation of TRMM rainfall data, the statistical criteria like mean error (ME), mean absolute error (MAE) and root mean square error (RMSE) were used and respectively were -1.34, 1.70 and 2.58 mm which accuracy is acceptable. Calibration of TRMM rainfall amounts were used by ground station's data and linear regression and its correlation coefficient was 0.69, which indicates the relative match of TRMM rainfall with ground station's data. Regression analysis using the F-statistic and Significant test of the regression line slope using the t test represents a significant match of TRMM rainfall with observed data at the 1% significance level. This research was conducted in the six-hour time scale. Therefore, it is suggested to the other researchers to evaluate The TRMM rainfall data on daily, monthly and annual scale and in other basins.