Validation efforts have been integral components of satellite missions in order to provide users with not only the desired products, but also the necessary data quality information. This strategy…

This federal contract award, valued at $520,034.00, was issued by the National Aeronautics and Space Administration (NASA) to Colorado State University (CSU) for a research project focused on improving the performance and applicability of the Goddard Profiling Algorithm (GPROF), a key component of NASA's Global Precipitation Measurement (GPM) mission.

The research aims to develop methods for generalizing validation statistics collected in specific regions, allowing GPROF performance to be predicted globally based on observed meteorological regimes. This is intended to enhance the usefulness of GPROF-derived precipitation estimates, particularly in areas lacking direct validation data. The project will examine regime-dependent errors at both instantaneous and seasonal timescales, with applications for NASA's Integrated Multi-satellitE Retrievals for GPM (IMERG) product as well as climate variability and trend studies. There is no set-aside designation mentioned for this award. As a leading public research university with expertise in areas like satellite data analysis, meteorology, and climate science, CSU is well-positioned to support this effort for NASA.

Validation efforts have been integral components of satellite missions in order to provide users with not only the desired products, but also the necessary data quality information. This strategy implicitly assumes that validation statistics are generally reproducible form one region to another so that a few validation efforts are enough to fully characterize products. Unfortunately, with passive microwave rainfall estimates, the information content is limited, and validation results often retain distinct regional differences. Well correlated and nearly unbiased results in the central united states, for instance, can become negatively biased with much poorer correlations when the same algorithm is applied over south korea. Here, we attack this problem specifically for the gprof algorithm, although some of the methods are likely generalizable to other rainfall sensors and products. Because regional problems are caused neither by random or systematic errors, but rather by regime dependent variations in the cloud systems, biases and uncertainties must also be framed in the regime framework. On an hourly basis, for instance, the diurnal cycle of precipitation is critical, but its importance diminishes as weeks and months are averaged. On a daily or synoptic scale, the general circulation and the relevant thermodynamic regimes are likely the biggest contributors to regional differences, while at large domain, seasonal scales, small, but persistent differences in the atmospheric forcings are likely to blame for regional biases such as those that have been observed between equatorial south america and africa. To overcome these limitations, we hope with this work, to begin generalizing validation statistics that have been collected for individual regions, to apply to observational and meteorological regimes that are applicable globally. If we can properly define these regimes such that gprof validation statistics are reproducible any time, and anywhere that the regime is observed, the validation paradigm can be extended to predict validation statistics globally as long as the defined regimes are well observed by some of gpms current validation networks. We focus here on regime-dependent errors on 1 , instantaneous and 5 seasonal (indexed by the central month) scales for their immediate applications to imerg in the case of 1 , daily precipitation, and for climate variability and trend studies in the case of 5 , seasonal grids. With this work, we hope to lay the foundation for predicting uncertainties at these two scales in regions without validation data arguably the regions in which satellite uncertainties are most critically needed. By relying on regimes (rather than locations) we hope to significantly expand the applicability of validation data that has been collected to date, thereby expanding the usability of this product, and the products who use gprof in the merged applications. While the approach of using regimes instead of geographic locations perhaps new to rainfall validation, IT is a concept that has long been used to better understand storm dynamics and the bedrock of many of the trmm and GPM field experiment planning. This provides some basic assurances that the method should prove useful for predicting uncertainties in remote locations, as long as their precipitation regimes overlap regimes that are well observed by our current networks of ground-based radars