*3.1. In-Situ Records*

The long term monthly precipitation data that were collected from 20 meteorological stations over the Punjab region were acquired from Pakistan Meteorological Department (PMD) (Figure 1). The in-situ records were investigated during the study periods from 1979 to 2017 and from 2003 to 2017 as reference data for the assessment of GB and SB precipitation products, respectively. The selection of the gauges depends on their maximum availability and completeness of data series. The GPPs were evaluated against the in-situ records on an annual, winter, and summer monsoon season. The winter and summer monsoon seasons are the primary drivers of the annual hydrological cycle with major influence of Asian monsoon (summer monsoon season) over the study region [28]. The GB and SB GPPs were developed while using quality-controlled procedures and algorithms. However, the stations data are generally considered as standard measurements for the evaluation of GPPs [37,38]. Moreover, the standard quality-controlled procedures were adopted for the better accuracy assessment of GB and SB GPPs. For quality assurance, station data series were thoroughly checked and outliers were fixed with neighboring gauge records [39]. Gaps were filled using a time-based interpolation approach [40]. Several post-processing techniques have been developed to identify the consistency and homogeneity's in the station records [41]. The double-mass curve was applied to observe the inhomogeneity in the station records [42]. The curve of all the stations showed a straight line with no evident break points, which confirms a temporal consistency and uniformity in the time series records.

After quality control process in the reference data series, the thin-plate splines (TPS) interpolation method was introduced to convert the station data into products relative spatial grid sizes for the assessment of GB and SB GPPs on spatial scale. The original TPS method is defined by [43], whereas [44] provides a detailed description of its application for different climate indicators. The TPS scheme is applicable and robust in regions where stations density is low and prior estimation of the spatial auto-covariance is not required [44]. The time series autocorrelation method was applied prior to estimate the trend significance, magnitude, and abrupt turning point to ensure the occurrence of significant autocorrelation in the reference data series [1]. The presence of significant autocorrelation in the reference data series could influence the outcome of Mann–Kendall (MK), Sen's Slope (SS), and Mann–Kendall abrupt change analysis [45]. Therefore, the significance of autocorrelation should be checked before applying the trend tests [46]. The analysis revealed no significant autocorrelation in the reference data series of annual, winter, and summer monsoon seasons. Therefore, the in-situ records are completely independent, and the MK test is applicable to the original gauge records. The detailed description for autocorrelation method is reported by [47].
