**1. Introduction**

Precipitation plays a crucial role in the Earth's hydrological cycle and is a fundamental input to a wide range of hydrological, meteorological, and climate model applications [1,2]. Thus, accurate estimation of the precipitation amount and pattern is vital for improved prediction of water-related processes as well as reducing uncertainties for effective water resource management practices [3,4]. To obtain precipitation amounts, ground-based measurements, i.e., rain gauges and weather radars, are considered a reliable source mainly at the local scale. At the regional and global scale, however, there are limitations for using ground-based measurements, particularly in most developing countries [5]. Radar networks are often available where there is a coverage by rain gauges. However, radars are subject to different errors and uncertainties, such as ground clutter, anomalous propagation, signal attenuation, beam blockage, and bright band contamination [6].

Rain gauges are limited in describing the spatial distribution of precipitation depending on the arrangement and density of the rain gauge network [7,8]. In order to spatially characterize precipitation, gauge measurements are transformed to a gridded precipitation dataset. This is carried out through interpolation of rain gauge measurements, using spatial interpolation and geo-statistical methods [9]. These may be prone to missing values, wind effects, insufficient numbers of rain gauges, and a sparse network, especially in less accessible mountainous and oceanic areas [4].

In view of the above, the spatial limitations, resolution, and coverage of ground-based measurements highlight the importance of satellite-based precipitation estimates at both the regional and global scale. Satellite-based precipitation estimates are also subject to uncertainties through cloud top reflectance, thermal radiance, infrequent satellite overpasses, and retrieval algorithm related to the nature of indirect measurement [10]. Therefore, a thorough validation of satellite precipitation data in any given area is necessary to achieve insight regarding is accuracy as well as identifying sources of errors to improve algorithms and satellite sensor development. Further, accuracy assessment taking into account the pros and cons of satellite precipitation estimates is imperative before using data in hydrological modeling in any given region [11,12]. Such findings help in selecting a supportive product for a special application under different circumstances [1].

Given the success of the Tropical Rainfall Measuring Mission (TRMM), the National Aeronautical and Space Administration (NASA) and Japan Aerospace Exploratory Agency (JAXA) launched a new generation Global Precipitation Measurement (GPM) mission in early 2014 to replace the TRMM mission [13]. The GPM mission is expected to compensate the limitations of TRMM precipitation products by providing higher resolution, larger spatial coverage, and more accurate global precipitation estimates [14]. The GPM precipitation algorithm, Integrated Multi-satellite Retrievals for GPM (IMERG), is based upon the experiences from the TRMM algorithm. As the spatiotemporal resolution and coverage of GPM have been extended beyond the TRMM resolution and coverage, the performance of the GPM IMERG products needs to be evaluated and validated globally.

Several studies have compared the GPM IMERG and TRMM products with ground-based measurements, i.e., rain gauge and weather radar [4,6,10,15–19], considering their hydrological applications [14,20–22]. Also, different GPM IMERG products regarding temporal resolutions have been evaluated considering various climatic and topographic conditions using various statistical measures across the world [5,23–30]. Although most of these studies confirmed the improvement of the IMERG products relative to those of the TRMM Multi-satellite Precipitation Analysis (TMPA), a more comprehensive investigation is still essential to better understand the IMERG performance in various regions of the world taking into consideration different products' versions and temporal resolution. Countries in the Middle East suffer from acute hydrometeorological data shortage, both in terms of quality and quantity [15], and Iran is not an exception. Rain gauges are sparse and unevenly distributed throughout the country, particularly in remote areas of the center and eastern areas. Delays in data processing and publishing for public access and scientific use and an absence of data sharing in many trans-boundary basins constitute a main shortcoming for ground-based precipitation data in the country [5]. To our knowledge, there are very few investigations of the IMERG products´ performance over Iran on a basin scale [5,10]. There are no comprehensive studies that investigate the performance of the IMERG product at the country level.

According to the above, the newly available IMERG products have not been thoroughly explored for Iran as a whole. The country covers different climatic, geographic, and topographic features, with respect to temporal and spatial particularities and different satellite products´ versions. This study aimed to provide a better understanding of the IMERG product's performance over the country and open a door to future studies regarding hydrological and hydrometeorological applications of these products at both the local and regional scale. Accordingly, we performed a comprehensive evaluation of the performance of IMERG products considering three time-latencies, IMERG-Early, IMERG-Late, and IMERG-Final, and two temporal resolutions, daily and monthly, based on eight criteria indices. We examined these criteria in view of spatial and temporal patterns related to features, such as elevation, slope, latitude, and longitude, over the entire Iran. Also, the statistical distributions of the precipitation products were compared to that of ground measurements for different seasons.
