Applicability of AgMERRA forcing dataset forgap-filling of in-situ meteorological observation, Case Study: Mashhad Plain

Consistency and transparency in climate data and methods facilitate comparisons across regions or between models in each of these assessments, particularly when market linkages between regions are emphasized (14 and 15). However, the density and quality of stationary climate data varies widely through space and time, with the best coverage in developed countries and less reliable coverage in the Tropics and Southern Hemisphere (15). So, several groups have collected these data and constructed harmonized, global gridded datasets at monthly resolution. However, these require weather generators synthesize daily resolution before they may be applied to crop models and are therefore likely to miss events that are important for the calibration and validation of agricultural models. Regional gridded observational networks have also been created (e.g., E-Obs in Europe, (8); APHRODITEin Asia, (21)), however many regions and variables are not covered by any such network and inter comparing sites between regions with different methodologies introduces inconsistencies (). Recently, AgMERRA climate forcing dataset provide daily, high-resolution, continuous, meteorological series over the 1980–2010 period designed for applications examining the agricultural impacts of climate variability and climate change. These datasets combine daily resolution data from retrospective analyses (the Modern-Era Retrospective Analysis for Research and Applications, MERRA) with in situ and remotelysensed observational datasets fortemperature, precipitation, and solar radiation, leading to substantial reductions in bias in comparisonto a network of 2324 agriculturalregion stations from the Hadley Integrated Surface Dataset (HadISD) (5).Therfore, this research was done in order to investigate the possibility of using AgMERRA climate forcing dataset to estimate missing data in in-situ daily temperature and precipitation observations in Mashhad plain.
The study area was Mashhad plain in KhorasanRazavi province, located in the northeast of Iran. Climate data corresponding to Mashhad plain extracted by means of geographical characteristics of Mashhad (for the 1980-2010 periods) and Golmakan (1987-2010 period) stations from AgMERRA dataset. The goodness of fit of AgMERRA climate forcing dataset was done by means of Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Mean Bias Error (MBE) tests and R2. The root mean-squared error (RMSE) is computed to measure the coincidence between measured and modelled values and Mean Bias Error (MBE) is simply to examine the overall model error.Furthermore, probability distribution function of observed daily data and AgMERRA data for both Golmakan and Mashhad stations calculated. Eventually, mean and variance of AgMERRA and in-situ data were calculated to have a more accurate comparison of simulated and observed data.
Results indicated that AgMERRA dataset has a good performance in estimating daily maximum and minimum temperature in Mashhad Plain. RMSE, MAE and MBE for daily precipitation illustrated a good performance of AgMERRA data. However, R2 value was 0.43 and 0.25 for Mashhad and Golmakan stations, respectively. Although the probability distribution function of daily maximum and minimum temperature and precipitation indicated the same trend for both studied stations, comparison of mean and variance of observed daily maximum and minimum temperature and precipitation and AgMERRA data for Mashhad and Golmakan stations showed different results. The difference between mean of AgMERRA and observed daily maximum temperature for Mashhadand Golmakan stations was 3.42 and 2.10 C°, respectively. It was 4.68 and 3.05 C° for minimum daily temperature for Mashhad and Golmakan, respectively, and the difference between mean of AgMERRA and observed daily precipitation was 0.06 and 0.28 mm.day-1 for Mashhad and Golmakan, respectively.
Discussion and This research showed that using AgMERRA climate forcing dataset could be a reliable tool to estimate missing data of in-situtemperature observations. Although the performance of AgMERRA dataset was good for daily precipitation, distribution of simulated precipitation compare with observed precipitation was different. Concerning AgMERRA precipitation data some points have to keep in mind that precipitation in arid and semi-arid regions tends to be more variable in time than in humid regions. In fact, the distinctive features of arid and semiarid regions affect precipitation modeling on a discrete event basis and a continuous basis (7, 10, 13).Results of this research illustrated the same trend and it revealed that AgMERRAdataset could not simulate the precipitation distribution in Mashhad plain. It seems that comparing AgMERRAprecipitation data with OPHRODITE dataset and other dataset can give us more accurate vision about AgMERRA dataset. Furthermore, it seems that it is needed to do more researches regarding investigation of performances of crop model results by using AgMERRA dataset as climate data input, because this dataset was released for agricultural application.