IPC 15 full program
Session 7
Future of precipitation products, education, and actionable early warning through partnerships.
Conveners: IPC Session & Karyn Tabor (Science Systems & Applications, Inc./NASA GSFC), Chris Funk (Climate Hazards Center)
June 17, 2026
Improving Hydrological Applications Through Enhanced Precipitation Estimation from Next Generation Gravity Missions Muhammad Usman Liaqat1, Stefania Camici1,Francesco Leopardi12, Jaime Gaona1,Luca Brocca1 1Research Institute for Geo-Hydrological Protection, National Research Council, Perugia, Italy. 2Department of Civil and Environmental Engineering, University of Perugia, Italy. Correspondence: muhammadusman.liaqat@cnr.it Satellite observations from the GRACE mission and GRACE-FO have significantly advanced the monitoring of terrestrial water storage (TWS) at regional to global scales. However, their coarse spatial and temporal resolution limits their applicability for representing precipitation variability required in hydrological applications. Emerging next-generation gravity mission concepts, including NGGM and MAGIC, aim to provide enhanced spatio-temporal observations of mass change, offering new opportunities for improving precipitation estimation and its use in hydrological modelling. The primary objective of this work to access how improving the spatial and temporal resolution of future gravity missions impacts precipitation estimation by developing a number of global synthetic experiments. The precipitation data used as forcing of ESM will be compared with the “true” precipitation for testing the reliability of the SM2RAIN approach (Brocca et al., 2014) using as input EWH data (in the past it was implemented by using surface soil moisture data). A suite of global synthetic experiments is conducted to evaluate multiple gravity mission configurations (GRACE-C-like, NGGM, and MAGIC), including both filtered and unfiltered scenarios, to assess the impact of improved spatial and temporal resolution on precipitation estimation. The global correlation analysis shows median and mean correlation coefficients of 0.67 and 0.63, respectively, indicating satisfactory performance of the EWH based SM2RAIN framework across most terrestrial regions. Stronger correlations are observed over Northern Hemisphere mid-latitudes, including Europe, northern Asia, and North America, reflecting robust performance in temperate climates, while reduced performance is evident in several tropical regions such as central Africa, parts of the Amazon Basin, and Southeast Asia. Ongoing work also includes validation against independent observational datasets, the results of which will be presented. The results of the study clearly highlight the added value of next generation gravity missions such as runoff simulation, soil moisture dynamics, and the monitoring of extreme events including floods and droughts. The proposed framework contributes to emerging high resolution hydrological modelling efforts and supports next generation earth system initiatives led by the European Commission (Destination Earth), ESA (Digital Twin Earth), and NASA (Earth System Digital Twin), aiming to improve early warning systems and water resource management under climate change. References Brocca, L., Ciabatta, L., Massari, C., Moramarco, T., Hahn, S., Hasenauer, S., Kidd, R., Dorigo, W., Wagner, W., & Levizzani, V. (2014). Soil as a natural rain gauge: Estimating global rainfall from satellite soil moisture data. Journal of Geophysical Research: Atmospheres, 119(9), 5128–5141.
Bringing GPM to the Classroom: An Educational Program for Teachers on Precipitation Measurement and Climate Awareness Authors: Lisa Milani (University of Maryland, College Park, Maryland, USA and NASA Goddard Space Flight Center, Greenbelt, Maryland, USA); Vasco Mantas (University of Coimbra, Coimbra, Portugal); Andrea Portier (Science Systems and Applications Inc., Lanham, Maryland, USA and NASA Goddard Space Flight Center, Greenbelt, Maryland, USA); Joana Aldino; Margarida Anastácio; Rui Andrade; Laura Besco; Ennio Cantoresi; Giulia Ciantra; Valeria Cuccato; Laura Insogna; Manuela Interlandi; Ilaria Piccioni; SIlvia Raisi; Elisabetta Ricci; Valeria Romani; Michela Zanella Corresponding author email: lmilani@umd.edu Abstract: NASA's Global Precipitation Measurement (GPM) Mission Mentorship Program includes an educational initiative targeting teachers and educators. Through this spin-off program, GPM specialists collaborate with educators to explore the water cycle, climate change, and precipitation within the context of the GPM mission. The primary objective is to equip teachers with information and resources that enable them to effectively transmit scientific knowledge to their students. Following three instructional sessions covering the water cycle, weather and climate, and the GPM mission, teachers work alongside GPM specialists to create hands-on projects for students. The initiative is designed to fit within local educational and institutional contexts by conforming to national educational standards, meeting teacher professional development needs, and incorporating subjects of regional and professional relevance. The initiative employs an interdisciplinary approach centered on precipitation study, ranging from direct measurement through rain gauges installed on school grounds to analytical work comparing field measurements with GPM satellite data. Student classroom dialogues about precipitation patterns and variability foster climate change awareness and connect this initiative to broader school-based sustainability programs. This paper outlines the program structure, summarizes the teacher-led practical projects, and provides links to classroom-ready resources available in Italian, Portuguese and English. The program continues to evolve, with plans to expand internationally, making educational materials accessible in multiple languages to minimize language barriers and enhance younger generations' access to Earth observation data.
Precipitation Forecast Skill Evaluation for the USACE FIRO Screening Process Authors: Eric J. Shearer (Coastal & Hydraulics Lab, US Army Engineer Research and Development Center); Rachel Weihs (Center for Western Weather & Water Extremes, UC San Diego); Elissa Yeates (Coastal & Hydraulics Lab, US Army Engineer Research and Development Center); Jay Cordeira (Center for Western Weather & Water Extremes, UC San Diego); Emily Slinskey (Center for Western Weather & Water Extremes, UC San Diego); F. Marty Ralph; Cary Talbot Corresponding author email: eric.j.shearer@erdc.dren.mil Abstract: This work showcases the systematic forecast skill evaluation procedure developed for the Forecast-Informed Reservoir Operations (FIRO) Screening Process; a cornerstone of the national FIRO Pathfinder Effort being pursued by the US Army Corps of Engineers (USACE). The objective of this evaluation is to provide a scalable, multi-stage assessment of precipitation forecast skill across contributing basins of USACE reservoirs. This skill assessment helps identify and prioritize USACE reservoirs that are promising candidates for FIRO viability assessments, which evaluate the potential to modernize operations by increasing flexibility in storage and release decisions using advanced weather forecasting. The procedure employs a phased approach to assess forecast skill. Stage A utilizes a two-pronged method for initial screening. The first prong is a qualitative assessment where site water managers evaluate whether current inflow forecasts are sufficiently accurate for operational decision-making. This is paired with a quantitative analysis of precipitation forecast skill at lead times relevant to water operations. A site is only eliminated on the basis of poor forecast skill if both assessments indicate insufficiency. Stage B conducts a more detailed assessment to produce a quantitative Forecast Skill Score. The score is a weighted composite of three distinct metrics evaluated at a site-specific required lead time: 1.Critical Success Index (CSI): Evaluates the skill of forecasting extreme precipitation events by comparing the number of correctly forecasted extreme rainfall events to the total number of extreme rainfall forecasts and observations. 2.Dry Forecast Failure Ratio (DFFR): Measures the ability to accurately forecast dry conditions at medium range (e.g. 5 days), a critical lead time for water retention decision making. 3.Forecast Error Tolerance (FET): Quantifies a reservoir’s physical ability to absorb under-forecasted inflow, assessing the reservoir system’s resilience to forecast errors. The Stage B quantitative assessment of forecast skill is brought together with additional assessments of site suitability for FIRO which inform recommendations about whether to pursue deeper inquiry to support operational change at the site. Scoring and recommendations are reviewed and finalized during Stage C, a collaborative dialogue with site water managers to incorporate expert judgment and site-specific nuances before determining final recommendations. The Screening Process establishes an objective and repeatable methodology for assessing potential FIRO readiness across a national portfolio of reservoirs. The Forecast Skill Score is a foundational component of the FIRO Suitability Index (FSI), which informs whether the potential benefits of FIRO outweigh the level of effort and inherent risks and guides final recommendations.
An Extreme Precipitation Event Database for the Contiguous U.S. Authors: Daniel Watters (University of Oklahoma); Pierre Kirstetter (University of Oklahoma; NOAA National Severe Storms Laboratory); Andy Newman (National Center for Atmospheric Research); Andy Wood (National Center for Atmospheric Research; Colorado School of Mines); Guoqiang Tang (Wuhan University) Corresponding author email: daniel.watters@ou.edu Abstract: Many extreme precipitation events have occurred over the Contiguous U.S. (CONUS) in recent years, however there is presently no database which quantifies these events and represents their spatiotemporal evolution. The objective of this work is to produce a novel CONUS extreme precipitation event database using the Multi-Radar Multi-Sensor (MRMS) system which has the potential to support hydrological applications, climate research, and risk and emergency management. The MRMS system provides a selection of high-resolution hydrometeorological products across CONUS (0.01°, 2-minute spatiotemporal resolution; 2014 - present), produced from polarimetric NEXRAD ground-based radars, rain gauges, and atmospheric model analyses. Our methodology for producing the extreme event database accounts for the high variability of precipitation processes in space and time, which MRMS enables due to its ability to track the spatiotemporal evolution of CONUS precipitation systems, including phase (rain, snow), type (convective, stratiform), rate, and other parameters. Furthermore, MRMS provides a stable and uniform observation system which is necessary to develop an event database that can bridge in-depth and consistent process characterization from sub-hourly to daily timescales. This database’s novelty is the automatic detection of extreme events at the daily scale and subsequent detailed diagnosis at 2-minute resolution. The event database is a key verification-oriented element of a NOAA project to develop a new multi-decadal probabilistic CONUS-wide 2-km resolution surface precipitation dataset. Key results include initial studies into extreme precipitation events with MRMS from the past decade (e.g., Texas Floods 2025, Oroville Dam 2017) including comparisons of key hydrological quantities (event accumulation, max daily intensity, and others) amongst events. Furthermore, studies into the automatic detection of extreme precipitation events in the MRMS record and uncertainty analyses are discussed. This work is directly relevant to IPC Session 3.2 ‘From Pixels to Decisions: Applications, Success Stories and Bottlenecks of Gridded Precipitation Products’ as it applies the gridded MRMS product to represent precipitation extremes.
Convection, Convergence, Cyclogenesis: Conjectures Authors: David L. Tweedy Corresponding author: dltdighton1@comcast.net A subset of extratropical cyclones that produce extreme quantities of convectively driven precipitation may intensify resulting from a nascent process originating from the convection itself. Denoted here as “convective instability of a nascent kind” (CINK), it is proposed to result from a reconfiguration of horizontal momentum rather than latent heating associated with CISK type instabilities. Downward momentum transport is seen as the primary source of “warm conveyor belt” (WCB) flows in the boundary layer. A feed-back process results when the WCB increases convective instability, perpetuates the momentum transfer, and thereby strengthens itself. Heavy total storm precipitation can occur along the WCB axis when CINK instability persists. An extratropical cyclone with extreme convection and total precipitation is discussed. A possible role of CINK instability in tropical cyclone intensification is proposed.
Rainfall detection at agricultural field level in Ghana using Sentinel-1 data Authors: Vincent Hoogelander (Delft University of Technology), Nick van de Giesen (Delft University of Technology), Rolf Hut (Delft University of Technology) Corresponding author: v.hoogelander@tudelft.nl In Ghana, large discrepancies between gridded satellite estimates and rain gauge data arise from spatial scale mismatches. Given the ground-data gap here, there is a need for alternative high-resolution satellite-based approaches to better understand rainfall variability at sub-pixel scale. In this study, we developed a data-driven method using high-resolution radar data from Sentinel-1 satellites to detect rainfall over small agricultural fields in Ghana. We assume that short-term changes in Sentinel-1 backscatter over agricultural fields during the start of the rainy season are primarily driven by variations in soil moisture. This implies that these data can be used directly to identify rainfall events over the field. We selected 32 agricultural fields, 16 around each of two rain gauges, and constructed a pixel-wise climatology of VV backscatter for each field. From this, we derived the relative wetness for individual pixels at each acquisition compared to this climatology. In the case that pixels within the field showed an increase in relative wetness between two consecutive Sentinel-1 acquisitions, this was likely caused by rainfall. The fields were manually evaluated against both ground-based and satellite rainfall data. A statistical test was taken to enable automated rainfall detection. Our results demonstrate that this approach can serve as a strong complement to coarser satellite rainfall products by providing insights into sub-pixel rainfall variability. The main source of uncertainty arises from rainfall timing relative to the revisit interval of Sentinel-1. This presentation highlights the potential of using Sentinel-1 to bridge the gap between point-based observations and coarse satellite rainfall products in regions with limited ground data.
Toward Improved Training Data for ML-Based Precipitation Retrievals Authors: Spencer R. Jones (Colorado State University, Fort Collins, Colorado, USA); Christian Kummerow (Colorado State University) Corresponding author email: spencer.jones@colostate.edu Abstract: Recent advances in machine learning (ML) and artificial intelligence techniques for satellite precipitation retrievals have catapulted these types of approaches into scientific prominence due to their relatively low computational cost, accuracy, and ability to regress very nonlinear relationships in high-dimensional space. The quality of the data used to train these models is likely becoming a limiting factor as performance becomes no longer limited by model architectural complexity and training schemes. Currently, the training data for most ML-based precipitation retrievals consists of simulated microwave brightness temperatures (Tbs) that are produced from individual plane-parallel model atmospheres—an assumption historically necessary due to computational limitations. However, some uncertainty arises in the training data due to these assumptions, as well as the ambiguity inherent in the observed Tbs due to overlapping fields of view (FOVs) and different horizontal resolutions of each channel. We have hypothesized that the training data can be improved by considering neighboring and overlapping FOVs simultaneously via a spatial approach that attempts to find the precipitation field that is consistent with both the observed Tbs and the spatial characteristics of precipitation fields as derived from high-resolution radar. This is expected to increase both validation performance and effective resolution of passive microwave retrievals.
Probabilistic Quantitative Precipitation Estimation With Space-Based Radars Authors: Pierre Kirstetter (University of Oklahoma); Nana Liu (University of Oklahoma) Corresponding author email: pierre-emmanuel.kirstetter@ou.edu Abstract: Progress in precipitation estimation is critical to advance weather and water studies and to characterize extreme events and associated natural hazards from local to global scales. Spaceborne weather radars on the Global Precipitation Measurement (GPM) and CloudSat satellites uniquely capture precipitation characteristics globally, serving as cornerstones for the global understanding of water fluxes. While Quantitative Precipitation Estimation (QPE) from these sensors is currently deterministic, applications and the comprehension of hydrometeorological processes require more than a single "best estimate" to effectively manage the intermittent, highly skewed distribution that characterizes precipitation. To advance the quantitative interpretation of spaceborne radar observations of precipitation, we propose the explicit estimation of uncertainty and extremes through probabilistic approaches. Probabilistic QPE leverages the Ground Validation Multi-Radar/Multi-Sensor (GV-MRMS; S-band) to demonstrate GPM Dual-frequency Precipitation Radar (DPR and Combined algorithm; Ku-Ka-band) and CloudSat (W-band) retrievals of QPE uncertainty and extremes. Probability distributions of precipitation rates are established using models quantifying the relation between sensor measurements, algorithm states, and the corresponding GV-MRMS reference precipitation. This approach integrates sources of error in spaceborne radar QPE and provides a framework for in-depth diagnostic of radar algorithms, especially when instruments sample raining scenes or processes challenging the algorithms’ assumptions. For practical applications, probabilistic QPE mitigates systematic biases from deterministic retrievals, quantifies uncertainty, and advances the monitoring of precipitation extremes through remote sensing. Probabilistic QPE opens perspectives for improved understanding of precipitation and its parameterization in spaceborne retrievals, estimation of precipitation at multiple scales, hydrological prediction, and risk monitoring.
Dependence of the Value of QPF on Basin Spatial Scale Authors: Witold Krajewski (University of Iowa); Felipe Quintero (University of Iowa); Nicolas Velasquez (Florida Institute of Technology) Corresponding author email: witold-krajewski@uiowa.edu Abstract: Quantitative Precipitation Forecasting benefits real-time streamflow forecasts by extending the lead time horizon. Uncertainties in QPF compromise these benefits. This study examined the performance of the short-term QPF product known as High Resolution Rapid Refresh (HRRR), used as the input to the National Water Model (NWM) for streamflow forecasting. The NWM streamflow output is examined at over 7,000 gauging stations operated by the United States Geological Survey. Results of several analyses are discussed. The first set of analyses compares HRRR QPF to the corresponding Quantitative Precipitation Estimation product known as Multi Radar Multi Sensor. Both the QPF and the QPE products represent hourly rainfall accumulations. The comparison is performed in the context of river basins with boundaries defined by the USGS gauging stations and ranging from ~10-1,000,000 km2. The second set of analyses focuses on streamflow error forecasted by the NWM and includes categorical evaluations of the ability of the QPF-driven NWM to detect floods, defined as discharge exceeding the Mean Annual Peak value. Other analyses examine error distributions. All analyses are conducted as function of lead time and spatial scale. Results demonstrate marginal benefit of the HRRR for streamflow forecast especially for basins smaller than 1,000 km2. For large basins, streamflow data assimilation plays a significant role in improving the forecasting skill.
Defining the Disconnect Between User and Researcher Needs for Informing the Next Generation of Multi- Satellite Precipitation Products: Lessons Learned from the GPM Era Authors: Andrea Portier (NASA GSFC/ Science Systems and Applications, Inc.); Sarah Ringerud (NASA GSFC); George Huffman (NASA GSFC); Emily Berndt (NASA MSFC) Corresponding author email: andrea.m.portier@nasa.gov Abstract: Since launch, the Global Precipitation Measurement (GPM) mission has been charged with maximizing societal benefit through the application of satellite precipitation data to support real-world decision-making. To meet this goal, the GPM Applications Team has led sustained engagement efforts through workshops, conferences, interviews, trainings, webinars, and educational outreach. These activities have not only facilitated the uptake of GPM data, but have systematically documented user needs, data challenges, and operational constraints across sectors. To date, these efforts have resulted in more than 140 captured case examples spanning weather forecasting, disaster response, public health, water resource management, ecological monitoring, and energy applications. A significant portion of this community relies on real-time products for time-sensitive decision-making with a particular focus on using NASA’s gridded precipitation product, the Integrated Multi-satellitE Retrievals for GPM (IMERG). The success of GPM applications can be largely attributed to the accessibility and timeliness of this data product. However, as the mission evolves and future multi-satellite products are planned, a critical gap has emerged. While stakeholder data needs are actively gathered, those needs may not be consistently translated into actionable technical requirements for algorithm developers. As a result, valuable user insights do not effectively inform data product design, retrieval development, or validation priorities. User needs are often articulated at a high level such as earlier warnings, improved situational awareness, better extreme event detection, yet algorithm developers require specific performance targets, uncertainty thresholds, latency requirements, spatial and temporal resolutions, and classification improvements to drive product evolution. The systematic capture of stakeholder needs represents a major achievement however, converting those documented insights into actionable guidance for algorithm developers continues to be a bottleneck in product evolution. This presentation is two-fold: it highlights representative use cases and lessons learned from the GPM applications era, and it examines how engagement practices can better elicit technically actionable requirements. We aim to foster dialogue on how stakeholder insights can more effectively inform algorithm development, enabling more decision-responsive precipitation products and enhancing the societal impact of future multi-satellite products.
Session summary
Back to full program.