Zadeh et al. (2026) A Global High-Resolution Precipitation Climate Record: PERSIANN-CCS-CDR Version 2.0

Identification

• Journal: Scientific Data
• Year: 2026
• Date: 2026-01-29
• Authors: Mehdi Rezaeian Zadeh, Phu Nguyen, Kuolin Hsu, Amir AghaKouchak, Tu Thanh Ung, Soroosh Sorooshian
• DOI: 10.1038/s41597-026-06625-5

Research Groups

• Center for Hydrometeorology and Remote Sensing, University of California, Irvine, USA

Short Summary

This paper introduces PERSIANN-CCS-CDR Version 2.0, a global high-resolution precipitation climate record, which resolves inconsistencies of its predecessor by providing two consistent sub-products (PERSIANN-CCS-CDR-B1 and PERSIANN-CCS-CDR-CPC) and evaluates their performance for extreme precipitation studies.

Objective

• To introduce PERSIANN-CCS-CDR Version 2.0, a new global high-resolution precipitation climate record.
• To address and resolve inconsistencies present in the previous version of PERSIANN-CCS-CDR, particularly those arising from the transition of input data sources.
• To evaluate the performance of the two new sub-products (PERSIANN-CCS-CDR-B1 and PERSIANN-CCS-CDR-CPC) as both a precipitation climate data record (using extreme indices) and a high-resolution precipitation record (during extreme events).
• To provide guidance for future users on the effective utilization of the new product based on its strengths and limitations.

Study Configuration

• Spatial Scale: Near-global coverage (60° S to 60° N, 180° W to 180° E), with a native spatial resolution of 0.04° by 0.04°. Intercomparison regions included the Contiguous United States (CONUS), Upper Mississippi basin, West Amazon Basin, and Mekong River Basin.
• Temporal Scale:
  • PERSIANN-CCS-CDR-B1: January 1983 to present.
  • PERSIANN-CCS-CDR-CPC: March 2000 to present.
  • Native temporal resolution of 3 hours (also available in 6-hourly, daily, monthly, and yearly aggregations).
  • Evaluation periods for climate indices: 1983–2000, 2000–2019, 2001–2024, and 2002–2023.
  • Evaluation periods for extreme events: Hurricane Michael (October 2018) and Upper Midwest severe weather outbreak (July 2024).

Methodology and Data

• Models used:
  • Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Cloud Classification System (PERSIANN-CCS) algorithm.
  • Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN) algorithm (for PERSIANN-CDR).
• Data sources:
  • Input for PERSIANN-CCS-CDR V2.0:
    • GridSat-B1 (NOAA’s Gridded Satellite B1 data): 3-hourly temporal resolution, 0.07° spatial resolution (for PERSIANN-CCS-CDR-B1).
    • CPC-4km (NOAA’s Climate Prediction Center globally merged IR product): 30-minute temporal resolution, 4 km spatial resolution (for PERSIANN-CCS-CDR-CPC).
  • Comparison/Reference Data:
    • NCEP/EMC 4KM Gridded Data (GRIB) STAGE IV Data (STAGE IV): Radar and gauge data, hourly, approximately 4 km resolution over CONUS.
    • PERSIANN-CDR: Near-global, 0.25° spatial resolution, daily.
    • International Best Track Archive for Climate Stewardship (IBTrACS): Best-track data for tropical cyclones (used for Hurricane Michael analysis).
    • Global Precipitation Climatology Project (GPCP) Monthly Analysis Product (used for bias correction of PERSIANN-CCS-CDR).
  • Ancillary Data: Shape files for region subsetting.
  • Evaluation Metrics: Extreme precipitation climate indices (PRCPTOT, R95PTOT, R99PTOT, SDII, R10TOT, R10, CDD, CWD), visual assessment, scatter plots, Root Mean Squared Error (RMSE), Correlation Coefficient (C.C.), percentile values, pixel count, Probability of Detection (POD), False Alarm Ratio (FAR), and Critical Success Index (CSI).

Main Results

• The previous PERSIANN-CCS-CDR exhibited significant inconsistencies around the year 2000, coinciding with the change in input data from GridSat-B1 to CPC-4km, which were clearly visible in extreme precipitation climate indices.
• PERSIANN-CCS-CDR Version 2.0 successfully resolves these inconsistencies by providing two separate, consistent sub-products: PERSIANN-CCS-CDR-B1 (since 1983) and PERSIANN-CCS-CDR-CPC (since March 2000).
• For the overlapping period (March 2000 to present), PERSIANN-CCS-CDR-CPC consistently outperforms PERSIANN-CCS-CDR-B1 when compared to STAGE IV, particularly in intensity-related extreme precipitation indices (SDII, R95PTOT, R99PTOT) and during extreme events (e.g., Hurricane Michael, Upper Midwest Flood).
• PERSIANN-CCS-CDR-B1, while skillful, shows some downsampling issues leading to coarser appearance and occasional unrealistically high precipitation values in some pixels, and its lower input frequency limits its ability to capture fast-evolving events.
• PERSIANN-CDR, despite its reliability at lower resolutions and for total precipitation, struggles to capture extreme precipitation values due to its 0.25° resolution and tends to overestimate light rain and event extent.
• Both PERSIANN-CCS-CDR-CPC and PERSIANN-CCS-CDR-B1 demonstrate strong agreement with STAGE IV for extreme events, but PERSIANN-CCS-CDR-CPC generally performs better in capturing the extent and intensity of these events.
• The PERSIANN algorithm has limitations in estimating precipitation from weakening hurricanes moving inland, as it primarily relies on cloud-top infrared temperature.
• Users are advised to prioritize PERSIANN-CCS-CDR-CPC for analyses requiring higher performance and accuracy, reserving PERSIANN-CCS-CDR-B1 for studies focused on extreme precipitation events prior to March 2000 or requiring a longer (over 40 years) high spatial-temporal resolution record.

Contributions

• Development and release of PERSIANN-CCS-CDR Version 2.0, a globally consistent, high spatial (0.04°) and temporal (3-hourly) resolution precipitation climate data record, addressing a critical need for monitoring extreme storms.
• Resolution of significant inconsistencies in the previous version of PERSIANN-CCS-CDR by introducing two distinct and consistent sub-products, enhancing data reliability for long-term climate studies.
• Comprehensive evaluation of the new sub-products against high-quality reference data (STAGE IV) and existing satellite products (PERSIANN-CDR) across various extreme precipitation indices and specific extreme events.
• Provision of clear, data-driven guidance for users on the optimal application of each sub-product, enabling informed decisions based on the trade-off between data record length and performance.
• The dataset is specifically designed and validated for high spatial-temporal resolution analyses, making it a valuable resource for hydrological and climate studies requiring detailed precipitation information.

Funding

• NOAA Climate Program Office’s Climate Observations and Monitoring Program (Grant NA23OAR4310443)
• NESDIS JPSS Program Office (Grant NA24NESX432C0001), administered by the Cooperative Institute for Satellite Earth System Studies (CISESS) at the University of Maryland/ESSIC
• NASA (Grants 80NSSC23K0136 and 80NSSC21K1668)
• Center for Western Weather and Water Extremes (CW3E) at the Scripps Institution of Oceanography through the AR Program (Grant 4600013361), sponsored by the California Department of Water Resources
• USACE FIRO III (Grant W912HZ2420001)
• NOAA (Grant ST133017CQ0058) with Riverside Technology, Inc.
• NVIDIA Academic Hardware Grant Program
• Google Cloud Research Credits program

Citation

@article{Zadeh2026Global,
  author = {Zadeh, Mehdi Rezaeian and Nguyen, Phu and Hsu, Kuolin and AghaKouchak, Amir and Ung, Tu Thanh and Sorooshian, Soroosh},
  title = {A Global High-Resolution Precipitation Climate Record: PERSIANN-CCS-CDR Version 2.0},
  journal = {Scientific Data},
  year = {2026},
  doi = {10.1038/s41597-026-06625-5},
  url = {https://doi.org/10.1038/s41597-026-06625-5}
}