Version 07 is the current version of the data set. Older versions will no longer be available and have been superseded by Version 07.The Integrated Multi-satellitE Retrievals for GPM (IMERG) IMERG is a NASA product estimating global surface precipitation rates at a high resolution of 0.1° every half-hour beginning 2000. It is part of the joint NASA-JAXA Global Precipitation Measurement (GPM) mission, using the GPM Core Observatory satellite as the standard to combine precipitation observations from an international constellation of satellites using advanced techniques. IMERG can be used for global-scale applications as well as over regions with sparse or no reliable surface observations. The fine spatial and temporal resolution of IMERG data allows them to be accumulated to the scale of the application for increased skill. IMERG has three Runs with varying latencies in response to a range of application needs: rapid-response applications (Early Run, 4-h latency), same/next-day applications (Late Run, 14-h latency), and post-real-time research (Final Run, 3.5-month latency). While IMERG strives for consistency and accuracy, satellite estimates of precipitation are expected to have lower skill over frozen surfaces, complex terrain, and coastal zones. As well, the changing GPM satellite constellation over time may introduce artifacts that affect studies focusing on multi-year changes.This dataset is the GPM Level 3 IMERG Late Daily 10 x 10 km (GPM_3IMERGDL) derived from the half-hourly GPM_3IMERGHHL. The derived result represents a Late expedited estimate of the daily mean precipitation rate in mm/day. The dataset is produced by first computing the mean precipitation rate in (mm/hour) in every grid cell, and then multiplying the result by 24. This minimizes the possible dry bias in versions before "07", in the simple daily totals for cells where less than 48 half-hourly observations are valid for the day. The latter under-sampling is very rare in the combined microwave-infrared (and rain gauge in the final) dataset, variable "precipitation", and appears in higher latitudes. Thus, in most cases users of global "precipitation" data will not notice any difference. This correction, however, is noticeable in the high-quality microwave retrieval, variable "MWprecipitation", where the occurrence of less than 48 valid half-hourly samples per day is very common. The counts of the valid half-hourly samples per day have always been provided as a separate variable, and users of daily data were advised to pay close attention to that variable and use it to calculate the correct precipitation daily rates. Starting with version "07", this is done in production to minimize possible misinterpretations of the data. The counts are still provided in the data, but they are only given to gauge the significance of the daily rates, and reconstruct the simple totals if someone wishes to do so. The latency of the derived Late daily product is a couple of minutes after the last granule of GPM_3IMERGHHL for the UTC data day is received at GES DISC. Since the target latency of GPM_3IMERGHHL is 14 hours, the daily should appear no earlier than 14 hours after the closure of the UTC day. For information on the original data (GPM_3IMERGHHL), please see the Documentation (Related URL). The daily mean rate (mm/day) is derived by first computing the mean precipitation rate (mm/hour) in a grid cell for the data day, and then multiplying the result by 24. Thus, for every grid cell we have Pdaily_mean = SUM{Pi * 1[Pi valid]} / Pdaily_cnt * 24, i=[1,Nf]Where:Pdaily_cnt = SUM{1[Pi valid]}Pi - half-hourly input, in (mm/hr)Nf - Number of half-hourly files per day, Nf=481[.] - Indicator function; 1 when Pi is valid, 0 otherwisePdaily_cnt - Number of valid retrievals in a grid cell per day.Grid cells for which Pdaily_cnt=0, are set to fill value in the Daily files.Note that Pi=0 is a valid value.Pdaily_cnt are provided in the data files as variables "precipitation_cnt" and "MWprecipitation_cnt", for correspondingly the microwave-IR-gauge and microwave-only retrievals. They are only given to gauge the significance of the daily rates, and reconstruct the simple totals if someone wishes to do so. There are various ways the daily error could be estimated from the source half-hourly random error (variable "randomError"). The daily error provided in the data files is calculated in a fashion similar to the daily mean precipitation rate. First, the mean of the squared half-hourly "randomError" for the day is computed, and the resulting (mm^2/hr) is converted to (mm^2/day). Finally, square root is taken to get the result in (mm/day):Perr_daily = { SUM{ (Perr_i)^2 * 1[Perr_i valid] ) } / Ncnt_err * 24}^0.5, i=[1,Nf]Ncnt_err = SUM( 1[Perr_i valid] )where:Perr_i - half-hourly input, "randomError", (mm/hr)Perr_daily - Magnitude of the daily error, (mm/day)Ncnt_err - Number of valid half-hour error estimatesAgain, the sum of squared "randomError" can be reconstructed, and other estimates can be derived using the available counts in the Daily files.

Version 05 is the current version of the data set. Older versions will no longer be available and have been superseded by Version 05. This dataset is the GPM Level 3 IMERG Late Daily 10 x 10 km (GPM_3IMERGDL) derived from the half-hourly GPM_3IMERGHHL. The derived result represents a Late expedited estimate of the daily accumulated precipitation. The dataset is produced at the NASA Goddard Earth Sciences (GES) Data and Information Services Center (DISC) by simply summing the valid precipitation retrievals for the day in GPM_3IMERGHHL and giving the result in (mm). The latency of the derived late daily product is a couple of minutes after the last granule of GPM_3IMERGHHL for the UTC data day is received at GES DISC. Since the target latency of GPM_3IMERGHHL is 12 hours, the daily should appear about 12 hours after the closure of the UTC day. For information on the original data (GPM_3IMERGHHL), please see the Documentation (Related URL).

The Integrated Multi-satellite Retrievals for GPM (IMERG) is the unified U.S. algorithm that provides the Day-1 multi-satellite precipitation product for the U.S. GPM team. The precipitation estimates from the various precipitation-relevant satellite passive microwave (PMW) sensors comprising the GPM constellation are computed using the 2014 version of the Goddard Profiling Algorithm (GPROF2014), then gridded, intercalibrated to the GPM Combined Instrument product, and combined into half-hourly 10 x 10 km fields.

These are provided to both the Climate Prediction Center (CPC) Morphing-Kalman Filter (CMORPH-KF) Lagrangian time interpolation scheme, and the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks Cloud Classification System (PERSIANN-CCS) re-calibration scheme. In parallel, CPC assembles the zenith-angle-corrected, intercalibrated "even-odd" geo-IR fields and forward them to PPS for use in the CMORPH-KF Lagrangian time interpolation scheme and the PERSIANN-CCS computation routines.

The PERSIANN-CCS estimates are computed (supported by an asynchronous re-calibration cycle) and sent to the CMORPH-KF Lagrangian time interpolation scheme. The CMORPH-KF Lagrangian time interpolation (supported by an asynchronous KF weights updating cycle) uses the PMW and IR estimates to create half-hourly estimates.

The IMERG system is run twice in near-real time: "Early" multi-satellite product, ~4 hr after observation time, and "Late" multi-satellite product ~12 hr after observation time. After that, the IMERG system does one "Final" run, once the monthly gauge analysis is received. The "Final" satellite-gauge product has a latency of ~2 months after the observation month.

The baseline for the near real-time Late and Late half-hour estimates is to be calibrated with climatological coefficients that vary by month and location, while in the Final post-real-time run the multi-satellite half-hour estimates are adjusted so that they sum to a monthly satellite-gauge combination. In all cases the output contains multiple fields that provide information on the input data, selected intermediate fields, and estimation quality.

The following describes the derivation in more details.

The daily accumulation is derived by summing valid retrievals in a grid cell for the data day. Since the 0.5-hourly source data are in mm/hr, a factor of 0.5 is applied to the sum. Thus, for every grid cell we have

Pdaily = 0.5 * SUM{Pi * 1[Pi valid]}, i=[1,Nf] Pdaily_cnt = SUM{1[Pi valid]}

where: Pdaily - Daily accumulation (mm) Pi - 0.5-hourly input, in (mm/hr) Nf - Number of 0.5-hourly files per day, Nf=48 1[.] - Indicator function; 1 when Pi is valid, 0 otherwise Pdaily_cnt - Number of valid retrievals in a grid cell per day.

Grid cells for which Pdaily_cnt=0, are set to fill value in the Daily files. Note that Pi=0 is a valid value.

On occasion, the 0.5-hourly source data have fill values for Pi in a very few grid cells. The total accumulation for such grid cells is still issued, inspite of the likelihood that thus resulting accumulation has a larger uncertainty in representing the "true" daily total. These events are easily detectable using "counts" variables that contain Pdaily_cnt, whereby users can screen out any grid cells for which Pdaily_cnt less than Nf.

There are various ways the accumulated daily error could be estimated from the source 0.5-hourly error. In this release, the daily error provided in the data files is calculated as follows. First, squared 0.5-hourly errors are summed, and then square root of the sum is taken. Similarly to the precipitation, a factor of 0.5 is finally applied:

Perr_daily = 0.5 * { SUM[ (Perr_i * 1[Perr_i valid])^2 ] }^0.5 , i=[1,Nf] Ncnt_err = SUM( 1[Perr_i valid] )

where: Perr_daily - Magnitude of the daily accumulated error power, (mm) Ncnt_err - The counts for the error variable

Thus computed Perr_daily represents the worst case scenario that assumes the error in the 0.5-hourly source data, which is given in mm/hr, is accumulating within the 0.5-hourly period of the source data and then during the day. These values, however, can easily be conveted to root mean square error estimate of the rainfall rate:

rms_err = { (Perr_daily/0.5) ^2 / Ncnt_err }^0.5 (mm/hr)

This estimate assumes that the error given in the 0.5-hourly files is representative of the error of the rainfall rate (mm/hr) within the 0.5-hour window of the files, and it is random throughout the day. Note, this should be interpreted as the error of the rainfall rate (mm/hr) for the day, not the daily accumulation.

The Integrated Multi-satellitE Retrievals for GPM (IMERG) is the unified U.S. algorithm that provides the multi-satellite precipitation product for the U.S. GPM team.

Version 06 is the current version of the data set. Older versions will no longer be available and have been superseded by Version 06.

This dataset is the GPM Level 3 IMERG Late Daily 10 x 10 km (GPM_3IMERGDL) derived from the half-hourly GPM_3IMERGHHL. The derived result represents a Late expedited estimate of the daily accumulated precipitation. The dataset is produced at the NASA Goddard Earth Sciences (GES) Data and Information Services Center (DISC) by simply summing the valid precipitation retrievals for the day in GPM_3IMERGHHL and giving the result in (mm). The latency of the derived late daily product is a couple of minutes after the last granule of GPM_3IMERGHHL for the UTC data day is received at GES DISC. Since the target latency of GPM_3IMERGHHL is 12 hours, the daily should appear about 12 hours after the closure of the UTC day. For information on the original data (GPM_3IMERGHHL), please see the Documentation (Related URL).

In the original IMERG algorithm, the precipitation estimates from the various precipitation-relevant satellite passive microwave (PMW) sensors comprising the GPM constellation are computed using the 2017 version of the Goddard Profiling Algorithm (GPROF2017), then gridded, intercalibrated to the GPM Combined Ku Radar-Radiometer Algorithm (CORRA) product, and merged into half-hourly 0.1°x0.1° (roughly 10x10 km) fields. Note that CORRA is adjusted to the monthly Global Precipitation Climatology Project (GPCP) Satellite-Gauge (SG) product over high-latitude ocean and tropical land to correct known biases.

The half-hourly intercalibrated merged PMW estimates are then input to both the Climate Prediction Center (CPC) Morphing-Kalman Filter (CMORPH-KF) Lagrangian time interpolation scheme and the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks Cloud Classification System (PERSIANN-CCS) re-calibration scheme. In parallel, CPC assembles the zenith-angle-corrected, intercalibrated merged geo-IR fields and forwards them to PPS for input to the PERSIANN-CCS algorithm (supported by an asynchronous re-calibration cycle) which are then input to the CMORPH-KF morphing (quasi-Lagrangian time interpolation) scheme.

The CMORPH-KF morphing (supported by an asynchronous KF weights updating cycle) uses the PMW and IR estimates to create half-hourly estimates. The motion vectors for the morphing are computed by maximizing the pattern correlation of successive hours of the vertically integrated vapor (TQV) provided by the Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2) and Goddard Earth Observing System model Version 5 (GEOS-5) Forward Processing (FP) for the post-real-time (Final) Run and the near-real-time (Early and Late) Runs, respectively. The KF uses the morphed data as the “forecast” and the IR estimates as the “observations”, with weighting that depends on the time interval(s) away from the microwave overpass time. The IR becomes important after about ±90 minutes away from the overpass time.

The IMERG system is run twice in near-real time:

"Early" multi-satellite product ~4 hr after observation time using only forward morphing and "Late" multi-satellite product ~14 hr after observation time, using both forward and backward morphing and once after the monthly gauge analysis is received:

"Final", satellite-gauge product ~3.5 months after the observation month, using both forward and backward morphing and including monthly gauge analyses.

Currently, the near-real-time Early and Late half-hourly estimates have no concluding calibration, while in the post-real-time Final Run the multi-satellite half-hourly estimates are adjusted so that they sum to the Final Run monthly satellite-gauge combination. In all cases the output contains multiple fields that provid