Global precipitation measurement mission products and services at the NASA GES DISC.
Launched on 27 February 2014, the NASA/JAXA Global Precipitation Measurement (GPM) mission core satellite and a constellation of international satellites not only greatly extend the spatial coverage from its predecessor [the Tropical Rainfall Measuring Mission (TRMM)], but also provide improved measurements of precipitation globally. For example, a new Ka-band precipitation radar and additional high-frequency channels in the microwave instrument have been added to the GPM core satellite for improving light rain and snowfall measurements. Furthermore, the Integrated Multi-satellite Retrievals for GPM (IMERG) have been significantly improved over the TRMM Multi-satellite Precipitation Analysis (TMPA) in terms of spatiotemporal resolution, spatial coverage, and more.
GPM datasets are available for research and applications at the NASA Goddard Earth Sciences (GES) Data and Information Services Center (DISC), home to the TRMM data archive as well. To new users, it can be a daunting task to locate a suitable GPM dataset. Even for experienced TRMM users, such activity can also be difficult due to many changes implemented because GPM datasets and services have been completely redesigned to accommodate changes in data structure, format, data volume, new technology, etc. Therefore, it is necessary to develop an overview document that guides users in locating datasets of interest and services that are suitable for their research and applications. Recognizing a very diverse user community consisting of users from different scientific disciplines, backgrounds, and countries with different levels of data downloading capabilities and Internet connectivity, the GES DISC has developed data services to facilitate GPM data access and exploration.
GPM DATA PRODUCTS. GPM data products at the GES DISC are organized and archived based on three product levels defined by the NASA Earth Observing System Data and Information System (EOSDIS): Level-1, Level-2, and Level-3. In some satellite missions, Level-1 products are subdivided into two categories: Level-1A and Level-1B. Level-1A is defined as, "Reconstructed, unprocessed instrument data at sensor's full resolution, time-referenced, and annotated with ancillary information, including radiometric and geometric calibration coefficients and geo-referencing parameters (e.g., platform ephemeris) computed and appended but not applied to Level-0 data." For Level-1B, it is defined as, "Level 1A data that have been processed to sensor units (not all instruments have Level-IB source data)." For GPM, an additional Level-1 category, Level-1C, has been added for common inter-calibrated microwave brightness temperature (Tc) products from GPM constellation satellites, which is necessary to ensure no systematic differences for multi-sensor and multi-satellite precipitation retrieval algorithms such as GPM IMERG.
Table 1 lists GPM Level-1 datasets. Besides Level-1 datasets from the GPM Microwave Imager (GMI) and the Dual-frequency Precipitation Radar (DPR) onboard the GPM core satellite, there are Level-1 datasets from other satellites in the GPM constellation. There is only one Level-IA dataset containing GMI unpacked packet data or raw data. There are 3 Level-1B datasets (1 from GMI and 2 from DPR). The remaining datasets are Level-1C as described above. Figure la is a sample of GMI Level-1C common calibrated brightness temperatures at 37 GHZ showing Tropical Cyclone Nanauk over the Arabian Sea on 11 June 2014. As seen in Table 1, Level-1 GPM datasets consist of reconstructed and unprocessed instrument data at sensor's resolution and therefore are best suitable for algorithm development and other special activities.
Level-2 datasets are defined as, "Derived geophysical variables at the same resolution and location as Level 1 source data." Table 1 lists GPM Level-2 datasets distributed at the GES DISC. It is seen that GPM Level-2 datasets include those from GPROF (the Goddard Profiling Algorithm) from the GPM constellation satellites, GPM DPR, and their combined datasets as well as latent heating products from DPR. Figure 1b is a sample of GMI Level-2 GPROF surface precipitation, showing Hurricane Arthur near the South Carolina and Georgia coasts on 3 July 2014. Since Level-2 GPM datasets contain geophysical variables at sensor's resolution, their usage is typically wider than Level-1 datasets; for example, Level-2 precipitation can be used in case studies, ground validation, model verification, etc.
Level-3 datasets are "Variables mapped on uniform space-time grid scales, usually with some completeness and consistency." Table 1 lists Level-3 half-hourly, daily, and monthly gridded datasets. Half-hourly datasets consist of IMERG products only. Daily datasets include daily gridded orbital mosaic (or ascending/descending for DPR) datasets from microwave sensors in the GPM constellation satellites and DPR as well as daily GMI and DPR combined datasets. Monthly datasets include all GPROF datasets from the constellation and one from IMERG. The most popular datasets are the multi-satellite, Multi-sensor, and multi-algorithm GPM IMERG products that include Early, Late, and Final Run. The Early and Late Run of IMERG consist of near-real-time monitoring products with climato-logical gauge calibration. For the Final Run, the Global Precipitation Climatology Center (GPCC) monthly monitoring gauge dataset is used for bias correction. The latencies from observation to public distribution are 6 h (Early Run), 18 h (Late Run), and 4 months (Final Run), respectively. Both spatial (0.1[degrees]) and temporal (half-hourly) resolutions of IMERG have been significantly improved compared to 0.25[degrees] and 3-hourly resolutions in TMPA. These improvements are important for hydrometeorological research and applications as well as other applications. Details about the IMERG datasets can be found in their technical documents. Figure 1c is an example from the half-hourly IMERG Final Run, showing heavy precipitation at 0300 GMT 15 June 2014 in the mid-western United States. Other monthly datasets derived from different satellites in the GPM constellation are useful for the understanding of uncertainties in global precipitation measurements.
GPM DATA SERVICES. GPM data services are crucial to facilitate data evaluation and access in order to maximize the use of datasets in research and applications. Precipitation dataset users are very diverse, consisting of college professors, researchers, operational forecasters, citizen scientists, high school students, etc. Some of them are first-time users of remote sensing products, and human-readable data formats such as ASCII are needed. The HDF5 data format is used in all GPM standard products. Special software and knowledge are required to handle such complex data structures. Format conversion is often needed for many users from different backgrounds. In addition, not all users need a global coverage, and a sub-setting capability is necessary to minimize data transfer and storage, which is particularly important for users from developing countries where Internet bandwidth can be very limited. On the other hand, hydrologic applications are closely associated with watersheds, and some applications use political boundaries such as states or counties. Having a GIS shape file capability is necessary to allow users downloading data only in an irregular shape area.
Mirador is a Google-based data search interface that allows searching, browsing, and retrieving of Earth science datasets at the GES DISC. Mirador will soon be replaced by a more powerful data service system called the Unified User Interface (UUI) to unify several existing user services and provide data, services, and information in one unified user interface. Without the UUI, users will have to visit different websites or portals for data sub-setting, visualization, document information, data recipes, etc. In short, the UUI will save users time and expedite data access.
The Simple Subset Wizard, or SSW (Fig. 2a), provides a simple and easy way to subset Level-3 and limited Level-2 datasets not only from the GES DISC but also from other NASA data centers such as the NASA Global Hydrology Resource Center, the NASA Langley Atmospheric Science Data Center, etc. The SSW contains a text input area for keyword search, a calendar for selecting beginning and ending times, and a spatial bounding box for choosing an area of interest (Fig. 2a). The SSW allows parameter sub-setting and format conversion (Fig. 2b). For example, the SSW can convert the original HDF5 format in the IMERG Final Run dataset to either NetCDF or ASCII (Fig. 2b). For those who are not familiar with HDF5 or NetCDF, ASCII is a user-friendly and human-readable format. After all these, the SSW generates a list of URLs that can be used for batch download with popular off-the shelf software packages such as Wget. Currently, the SSW provides data sub-setting and format conversion services for all GPM Level-3 products listed in Table 1--including the popular IMERG datasets-except the latent heating datasets.
The Open Source Project for a Network Data Access Protocol (OPeNDAP) provides interoperability and remote access to individual variables within datasets in a form usable by many tools, including IDV, McIDAS-V, Panoply, Ferret, and GrADS. Format conversion can be achieved through the OPeNDAP, and available formats are ASCII, NetCDF 3, NetCDF 4, and binary. In addition to interoperability, the OPeNDAP is very useful for supporting operational activities because users can write a script to automatically pull data from the OPeNDAP on a fixed schedule. All datasets listed in Table 1 can be accessed through the OPeNDAP.
GPM DATA EXPLORATION. Giovanni (an acronym for the Geo-spatial Interactive Online Visualization and Analysis Infrastructure) is an online tool, developed at the GES DISC, to facilitate access, evaluation, and exploration of Earth science datasets. All IMERG datasets can be easily visualized and analyzed online with Giovanni without the need to download data and software. For novices, using satellite remote sensing datasets can be a daunting task, and numerous issues can be encountered in data processing, such as data format, data structure, data volume, Internet connectivity or bandwidth, etc. Moving a large amount of remote sensing data over the Internet can be time consuming and problematic for countries with low bandwidth and unreliable Internet connections. Sending a graphic result or time series in ASCII instead, other than the entire dataset, can make a significant difference to users in those countries. Nonetheless, online tools like Giovanni can provide a convenient way to bridge GPM data and users.
Recently, Giovanni has been completely redesigned due to an increasing demand for integrated analysis and visualization of a large collection of Earth science datasets at the GES DISC and other NASA data centers. Meanwhile, Giovanni evolves with modern software technologies and development to make it more user-friendly and increase its performance for data exploration. Giovanni contains only one landing page (Fig. 3a). Keyword and facet search capabilities (Fig. 3a) make searching a large amount of datasets a simple process. Due to a large amount of variables (more than -1,400 as of this writing) in Giovanni, the list of search results can sometimes be very long and difficult for users to locate a variable of interest. Faceting makes picking a dataset easy. For example, if one looks for calibrated precipitation in the monthly IMERG Final Run dataset, a search for "precipitation" in Giovanni returns a list of 102 variables from TRMM, GPM, MERRA (Modern Era Retrospective-Analysis for Research and Applications), NLDAS (Global Land Data Assimilation System), etc. By choosing GPM from the facet list, the list is shortened to 17 variables, and after clicking on "monthly" in Temporal Resolutions, only 4 variables are available and they all belong to the IMERG monthly product. Of course, one can simply search "IMERG Monthly" without doing any filtering work. IMERG Early, Late, and Final Run are available in Giovanni. For example, a trio of typhoons in the Western Pacific is shown in the rainfall intensity from the IMERG Early Run at 0200 GMT 7 July 2015 (Fig. 4a). For the time being, facets in Giovanni contain disciplines, measurements, platform/instrument, spatial resolutions, temporal resolutions, wavelengths, depths, special features, and portals. For users who are familiar with the TRMM Online Visualization and Analysis System (TOVAS), simply type in "TOVAS" in the search box and it retrieves all TRMM- and GPM-related variables in Giovanni.
Table 2 lists Giovanni plot types and formats available for data downloads. New functions are still being added to Giovanni. In the map group (Table 2), the "Accumulated" function allows users to generate an accumulated precipitation map from either a rectangular box or a shape (countries, major watersheds, and states in the United States). A sample is presented in Fig. 3b. The "User-Defined-Climatology" allows defining custom climatology for a user-defined time period so one can compare climatologies with different time lengths. In the comparison group, the interactive scatter map allows picking a point in a scatter-plot to show its geographic location, which can be useful for investigating unusual points in a scatter-plot. In the time series group, the "Seasonal" option allows users to choose one or multiple seasons and plot the time series. To compare with a gauge-based time series, users can input the latitude and longitude of a gauge location in the landing page and use the time series function to obtain the plot and the ASCII data in CSV (Comma Separated Values) for comparison. To create maps and time series for an irregular shape such as countries, states, and watersheds, one can click on the "Show Map" button and select a shape. Figure 3b is a sample rainfall map, showing heavy rainfall in the capital region of Tokyo due to the passages of Super Typhoons Phanfone and Vongfong in October 2014.
Giovanni output can be downloaded as well. Users can download images in the GeoTIFF, KMZ, and PNG formats. Digital map data can be downloaded as NetCDF, which is a very common format in many scientific communities and is easily imported into GIS software packages such as ArcGIS. Nongridded 2D data from time series, zonal mean, etc., can be downloaded as ASCII CSV, which can be imported into Microsoft Excel for further analysis.
Giovanni allows users to explore other precipitation datasets such as those from TRMM, MERRA, NLDAS, etc. For precipitation, it is well known that different units are used in different disciplines; for example, the units in IMERG (mm [hr.sup.-1]) and MERRA (kg [m.sup.2][s.sup.-1]) monthly precipitation products are different. In addition, their grid structures are different. Unit conversion and re-gridding algorithms are available in Giovanni, making the comparison of these monthly products possible. Figure 4b shows a difference map between the TMPA and IMERG Final Run monthly datasets in July 2014. It is seen that the IMERG precipitation is in general higher than that of TMPA over land and lower over oceans for July 2014 (Fig. 4b). Figure 4c is a scatter-plot between the two variables, showing a close relationship.
GPM DATA APPLICATIONS. Societal impacts are an important component of the GPM mission. Since the TRMM era, the GES DISC has developed data services to support domestic and international users in their precipitation-related applications. Based on user's reports, these applications include flood/drought monitoring activities, crop monitoring, disease studies/monitoring, hurricane watch, insurance industries, etc. In addition to the data services that have been mentioned above, the GES DISC is working closely with U.S. federal agencies such as the United States Department of Agriculture (USDA) Foreign Agriculture Service (FAS) to develop data services and support their worldwide operation. For example, the near-real-time TMPA 10-day product and anomaly have been in operation in the USDA Crop Explorer since TRMM. As soon as the retrospective processing of the IMERG data in the TRMM era is finished, we will work with USDA FAS to replace the TMPA product with the higher spatial resolution (0.1[degrees]) IMERG near-real-time product.
FUTURE PLANS. Future plans consist of two areas: value-added products and services. Value-added products are being developed to facilitate data access and scientific investigation activities. For example, not all users need half-hourly IMERG products, and daily products are sufficient to meet their requirements. Such daily products are available now. We will work closely with users and algorithm developers to develop additional value-added precipitation products. As for data services, more can be added to the existing services. Sub-setting Level-1 and Level-2 datasets is needed to avoid downloading unwanted data outside an area of interest. Although the OPeNDAP can perform such tasks, it is not as straightforward as the SSW, where a web interface is available for collecting user's input and generating a list of URLs for batch data download. Visualization and analysis of GPM Level-1 and Level-2 datasets at sensor's resolution in Giovanni are helpful for case studies, dataset evaluation, and algorithm development. For Level-3 products, custom datasets are necessary for those who use different grid structures, spatial and temporal resolutions, and projections in their activities. The GES DISC is also home to many NASA satellite missions or projects. Capabilities to integrate, analyze, and visualize datasets from other satellite missions or projects such as CloudSat, TRMM, etc., are also necessary for data exploration and scientific discovery. Event-based sub-setting services can save time because users do not need to use different sub-setters (if available) for obtaining data subsets, which is particularly useful for case studies.
ACKNOWLEDGMENTS. The authors thank GPM science team members--in particular, George Huffman--and many users for providing comments and suggestions during the dataset and service development at the GES DISC. Thanks also to Andrey Savtchenko for his contribution and to three anonymous reviewers for their comments and suggestions that have significantly improved and strengthened the manuscript. GPM datasets are processed and provided by the Precipitation Processing System (PPS) that also distributes GPM data and provides services.
FOR FURTHER READING
Garstang, M., and C. D. Kummerow, 2000: The Joanne Simpson Special Issue on the Tropical Rainfall Measuring Mission (TRMM). J. Appl. Meteor., 39, 1961.
Hou, A. Y., and Coauthors, 2014: The Global Precipitation Measurement Mission. Bull. Amer. Meteor. Soc., 95, 701-722, doi:10.1175/BAMS-D-13-00164.1.
Liu, Z., D. Ostrenga, W. Teng, and S, Kempler, 2012: Tropical Rainfall Measuring Mission (TRMM) precipitation data services for research and applications. Bull. Amer. Meteor. Soc., 93, 1317-1325.
NASA, 2016: Data Processing Levels. [Available online at https://science.nasa.gov/earth-science/earth -science-data/data-processing-levels-for-eosdis -data-products/.]
--, 2016: GPM IMERG data in Giovanni. [Available online at https://giovanni.sci.gsfc.nasa.gov/giovanni /#service=TmAvMp&starttime=&endtime=&bbox =-180,-90,180,90&dataKeyword=imerg.]
NASA GES DISC, 2016: GPM and TRMM data access through Mirador. [Available online at https://mirador .gsfc.nasa.gov/.]
--, 2016: GPM data access through UUI. [Available online at https://disc.sci.gsfc.nasa.gov/uui /datasets?keywords=GPM.]
--, 2016: OPeNDAP access. [Available online at https://gpm1.gesdisc.eosdis.nasa.gov/opendap/.]
--, 2016: Simple Subset Wizard. [Available online at https://disc.sci.gsfc.nasa.gov/SSW/#keywords=GPM.] NASA Goddard Space Flight Center, 2016: GPM documents. [Available online at https://pps.gsfc.nasa.gov /GPMprelimdocs.html.]
--, 2016: NASA Precipitation Measurement Missions web portal. [Available online at https://pmm.nasa.gov.] U.S.D.A., 2016: Crop Explorer. [Available online at www .pecad.fas.usda.gov/cropexplorer/.]
AFFILIATIONS: Liu--NASA Goddard Earth Sciences (GES) Data and Information Services Center (DISC), Greenbelt, Maryland, and Center for Spatial Information Science and Systems (CSISS), George Mason University, Fairfax, Virginia; Ostrenga, Deshong, and Macritchie--NASA Goddard Earth Sciences (GES) Data and Information Services Center (DISC) Greenbelt, Maryland, and Adnet Systems, Inc., Bethesda, Maryland; Vollmer and Kempler--NASA Goddard Earth Sciences (GES) Data and Information Services Center (DISC) Greenbelt, Maryland; Greene--NASA Goddard Earth Sciences (GES) Data and Information Services Center (DISC) Greenbelt, Maryland, and Wyle Information Systems, LLC, McLean, Virginia
CORRESPONDING AUTHOR E-MAIL: Zhong Liu, firstname.lastname@example.org
Caption: FIG. 1. Samples of GPM datasets at different levels, a) Level-1C GMI common calibrated brightness temperatures at 37 GHZ showing Tropical Cyclone Nanauk over the Arabian Sea on 11 Jun 2014. b) Level-2 GMI GPROF surface precipitation showing Hurricane Arthur near the South Carolina and Georgia coasts on 3 Jul 2014. c) Level-3 half-hourly IMERG Final precipitation showing heavy precipitation at 0300 GMT 15 Jun 2014 in the midwestern United States.
Caption: FIG. 2. The Simple Subset Wizard (SSW) provides a simple way to subset Level-3 and Level-2 (limited) GPM datasets not only from the GES DISC, but also other NASA data centers, a) The landing page of SSW. b) Sample output showing different options for subsetting data.
Caption: FIG. 3. The GES DISC Giovanni allows visualization and easy access to IMERG data, a) A screenshot of the Giovanni landing page. Features such as keyword and facets make dataset search simple, b) A sample rainfall map from IMERG Final Run showing heavy rainfall (in mm) in the capital region of Tokyo due to the passages of Super Typhoons Phanfone and Vongfong in October 2014.
Caption: FIG. 4. Samples of IMERG in Giovanni, a) Rainfall intensity (mm [hr.sup.-1]) from the IMERG Early Run at 0200 GMT 7 Jul 2015 showing a trio of typhoons in the Western Pacific, b) A monthly precipitation difference map (mm [hr.sup.-1]) between 3B43 and IMERG Final Run for July 2014, and c) their scatter-plot.
Table 1. GPM datasets at the GES DISC. Level-1 Dataset (reconstructed and unprocessed observations at sensor's resolution) GMI unpacked packet data GMI brightness temperatures GMI common calibrated brightness temperatures collocated Common calibrated brightness temperatures from the constellation of satellites (DMSP F/6-/9 SSMIS, GCOM-WI AMSR2, GMI, MetOp-A MHS, MetOp-B MHS, MTI SAPHIR, NOAA-18 MHS, NOAA-19 MHS, TRMM TMI, NPP ATMS) GPM DPR Level-1B Ku-band received power GPM DPR Level-1B Ka-band received power Level-2 Dataset (derived geophysical variables at sensor's resolution) GPM DPR environment GPM DPR Ku precipitation GPM DPR Ka precipitation Radiometer profiling from the constellation of satellites (DMSP FI6-19 SSMIS, GCOM-WI AMSR2, GMI, MetOp-A MHS, MetOp-B MHS, MTI SAPHIR, NOAA-18 MHS, NOAA-19 MHS, TRMM TMI, NPP ATMS) GPM DPR and GMI combined precipitation GPM DPR convective stratiform heating GPM DPR spectral latent heating Level-3 Dataset (variables on uniform space-time grid) GPM DPR daily (ascending, descending) GPM DPR daily and monthly precipitation profiles GPM DPR, GMI combined daily and monthly precipitation Daily and monthly GPROF profiling from the constellation of satellites (DMSP F-6-9 SSMIS, GCOM-WI AMSR2, GMI, MetOp-A MHS, MetOp-B MHS, MTI SAPHIR, NOAA-18 MHS, NOAA-19 MHS, TRMM TMI, NPP ATMS) IMERG half-hourly and daily (early run, late run, and final run) IMERG monthly (final run) GPM DPR daily and monthly convective stratiform heating and spectral latent heating Table 2. Giovanni plot types. File formats for downloads are PNG, GeoTIFF, KMZ, NetCDF, and ASCII (nongridded 2D data only). Maps Comparisons Time Series Time-averaged map Correlation map Area-averaged differences Animation Scatter-plot, area- Area-averaged averaged (static) Difference of time- Scatter-plot Seasonal averaged maps (interactive) (inter-annual) Accumulated map Scatter-plot (static) Hovmoller, longitude-averaged User-defined Scatter-plot, time- Hovmoller, climatology averaged (interactive) latitude-averaged Maps Vertical Plots Miscellaneous Time-averaged map Cross-section map, Zonal mean latitude-pressure Animation Cross-section map, Histogram longitude-pressure Difference of time- Cross-section map, averaged maps time-pressure Accumulated map Vertical profile User-defined climatology
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||NOWCAST: DATA ACCESS|
|Author:||Liu, Z.; Ostrenga, D.; Vollmer, B.; Deshong, B.; Macritchie, K.; Greene, M.; Kempler, S.|
|Publication:||Bulletin of the American Meteorological Society|
|Date:||Mar 1, 2017|
|Next Article:||The Abisko Polar Prediction School.|