Author: okelley

CEOSR personnel working at NASA’s Precipitation Processing System (PPS) have generated 27 years of global rainfall imagery that is now available in NASA Worldview, which is NASA’s online archive of Earth science imagery.

NASA’s Worldview website now allows you to explore global estimates of rainfall and snowfall from 1998 to the present at 30-minute intervals. Researchers and application developers have been using this dataset since 2014 (Portier 2024; Portier et al. 2023), but now it is also available as images in Worldview.

This precipitation dataset is called IMERG, which stands for Integrated Multi-satellitE Retrievals for Global precipitation measurement (GPM). Over the years, a team of NASA scientists have improved the IMERG algorithm’s accuracy over land and ocean (Huffman et al. 2023a,b,c).

Worldview is now being updated every hour with the most recent global IMERG estimates. Worldview is a user-friendly website for visualizing many of NASA’s Earth science datasets. The image below shows the precipitation falling from Hurricane Helene shortly before it made landfall in Florida in September 2024 (West et al. 2024). In the image, you can see how the IMERG estimates cover land and also ocean where there are no rain gauges or ground radars.

NASA runs the IMERG algorithm three times for each 30-minute period. The precipitation estimates from the first run become available just 5 hours after satellite observations are made, and that run is called Early IMERG. This near real-time dataset is useful for forecasting floods and monitoring large storm systems. A second run, called Late IMERG, occurs about 15 hours after observations are made. About four months later, the IMERG algorithm is run one last time when more observations and calibration datasets are available as inputs. This last run is called Final IMERG. Final IMERG is useful for studying droughts and other seasonal deviations from normal conditions. It can also be used for post-season reexamination of major events such as hurricanes. The Worldview website displays Final IMERG when available and Early IMERG otherwise.

Worldview for Beginners

If you haven’t used NASA’s Worldview website before, it is easy to learn how. Start by visiting https://worldview.earthdata.nasa.gov. There is no App to download, no user registration, and no charge.

When you visit Worldview, it initially loads satellite imagery of the most recent observations of the Earth’s surface and clouds. Precipitation is not initially displayed.

It takes two steps to display IMERG precipitation in Worldview. The first step is to add to your session the Worldview layer that contains 30-minute IMERG precipitation estimates. To do so, click the Add Layer button in the lower left corner. In the panel that pops up, type “imerg” in the search field. In the search results, check the box next to the entry for 30-minute IMERG. To dismiss the Add Layer panel, click the “X” symbol in the upper right.

The second step is to move a few hours into the past.  To do so, you can move the time slider a little to the left in the timeline along the bottom of the Worldview window. Alternatively, you can use the arrow icons to move the clock display back a few hours. This time adjustment is necessary because IMERG precipitation estimates are created approximately 5 hours after satellite observations are made and it takes another hour or so to convert the data into Worldview imagery.

The 30-minute IMERG imagery should now be visible with green-yellow-red colors for rainfall estimates and cyan-blue-purple colors for snowfall estimates. To adjust the map’s zoom level, use the plus and minus buttons in the Worldview window or the scroll wheel on your mouse if you have one. To change the map’s location, drag the map with your mouse. By default, Worldview shows coastlines. For more detailed political boundaries and place labels, turn on optional layers. These optional layers can be found at the top of the layer list on the left side of Worldview.

The IMERG algorithm estimates the precipitation rate at the middle of each 30-minute period, i.e., 15 minute and 45 minute after the hour. The units of precipitation rate are the millimeters of rain-equivalent depth that would accumulate in an hour if the observed precipitation rate (rainfall or snowfall) were to persist for a full hour. If the precipitation is frozen, then the value that IMERG reports is the depth of liquid water that would result if the frozen precipitation were melted before its depth is measured. The liquid-equivalent depth of snowfall is relevant to the water storage in snowpack, which influences how much water will become available to plants once the snow melts. Because fresh snow contains air (i.e., it’s fluffy), about 4 to 20 inches of snowfall will melt down to just an inch of liquid water (Baxter et al. 2005).  IMERG states the liquid-equivalent depth of snowfall, but most weather reports state the depth of snowfall in its original, frozen form. 

One advantage to viewing IMERG precipitation imagery within Worldview is that one can hover over an image in Worldview to obtain a numerical range for the precipitation rate represented by that pixel of the image. The numerical range is displayed in the color bar on the left side of the browser window. This color-table-interrogation feature is available only if you are using a device with a mouse that can hover, such as a laptop or desktop but not a mobile device.

If you encounter any trouble downloading IMERG data, contact the GPM project using the contact page.

Credits: Story and visualization by Owen Kelley.

References

Baxter et al., 2005: A Climatology of Snow-to-Liquid Ratio for the Contiguous United States. Weather and Forestcasting, 20, 729–744, Figs. 3 and 5.

Huffman, G. J., D. T. Bolvin, D. Braithwaite, K. Hsu, R. Joyce, C. Kidd, E. J. Nelkin, S. Sorooshian, J. Tan, and P. Xie, 2023a: NASA Global Precipitation Measurement (GPM) Integrated Multi-satellitE Retrievals for GPM (IMERG) Version 07, Algorithm Theoretical Basis Document (ATBD). https://arthurhou.pps.eosdis.nasa.gov/Documents/IMERG_V07_ATBD_final.pdf.

Huffman, G. J., D. T. Bolvin, R. Joyce, O. A. Kelley, E. J. Nelkin, J. Tan, D. C. Watters, and B. J. West, 2023b: Integrated Multi-satellitE Retrievals for GPM (IMERG) Technical Documentation. white paper, https://arthurhou.pps.eosdis.nasa.gov/Documents/IMERG_TechnicalDocumentation_final.pdf.

Huffman, G. J., D. T. Bolvin, R. Joyce, O. A. Kelley, E. J. Nelkin, J. Tan, D. C. Watters, and B. J. West, 2023c: IMERG V07 Release Notes. white paper, https://arthurhou.pps.eosdis.nasa.gov/Documents/IMERG_V07_ReleaseNotes_final.pdf.

Portier, A., 2024: Using GPM Precipitation Data to Help Inform Decisions and Policies Throughout the Globe. GPM 10-in-10 Webinary Series: Application, NASA, https://www.youtube.com/watch?v=NPg_2H-XxR4.

Portier, A., D. Kirschbaum, M. Gebremichael, E. Kemp, S. Kumar, I. Llabres, E. Snodgrass, and J. Wegiel, 2023: NASA’s Global Precipitation Measurement Mission: Leveraging Stakeholder Engagement & Applications Activities to Inform Decision-making. Remote Sensing Applications: Society and Environment, https://doi.org/10.1016/j.rsase.2022.100853.

West, J., S. Lang, and J. Reed, 2024: Powerful Hurricane Helene Makes Landfall in Florida’s Big Bend. GPM website, https://gpm.nasa.gov/applications/weather/news/powerful-hurricane-helene-makes-landfall-floridas-big-bend.

• *The programs and services offered by George Mason University are open to all who seek them. George Mason does not discriminate on the basis of race, color, religion, ethnic national origin (including shared ancestry and/or ethnic characteristics), sex, disability, military status (including veteran status), sexual orientation, gender identity, gender expression, age, marital status, pregnancy status, genetic information, or any other characteristic protected by law. After an initial review of its policies and practices, the university affirms its commitment to meet all federal mandates as articulated in federal law, as well as recent executive orders and federal agency directives.

On Monday evening, Sept. 23, 2024, an unusual hurricane struck southern Mexico, dumping 20 inches of rainfall along the coast according to estimates from NASA’s IMERG precipitation product.  The following image shows these rainfall estimates.

Observations from an international constellation of satellites were the starting point for calculating these estimates, and the data confirm that forecasters in prior days were right to warn of possible heavy rainfall. Forecasters did have trouble predicting the storm’s track and intensity, though. In the rural area of southwestern Mexico where Hurricane John struck, there are few rain gauges that provide public data in real-time, so having calibrated multi-satellite near real-time precipitation estimates like IMERG available is particularly valuable for understanding this storm as it unfolded. 

The contours in the image cover the areas where the IMERG algorithm estimated that 5 inches (blue) to 36 inches (bright red) of rainfall fell during a four-and-a-half-day period ending on Friday, Sept. 27, at 3 a.m. CST (0900 UTC). The dashed yellow line in the image shows the U.S. National Hurricane Center’s estimate of the path taken by the storm’s low-pressure center.

According to NOAA hurricane track data, approximately once a year during the past 50 years, one major hurricane over the eastern Pacific has passed within 800 miles of this part of southern Mexico. Almost all of these hurricanes moved roughly west over the ocean, parallel to the coast. Before Hurricane John this year, only two other times during the past 50 years has a major hurricane made landfall within 150 miles of where Hurricane John hit Mexico on Monday. On Sept. 23, John made landfall near the small, coastal village of Punta Maldonado, which is located about 100 miles east of the town of Acapulco. The two other hurricanes that had made landfall in the same region in prior years were Hurricane Pauline in 1997 and Hurricane Otis in 2023.

John was just a 40-knot tropical storm at 3 a.m. CST on Monday, and forecasters had only expected it to become a slightly stronger tropical storm during the following 24 hours before making landfall and dissipating. Instead, John experienced extremely rapid intensification to become a 105-knot Category 3 hurricane on the Saffir-Simpson scale. This intensification occurred during a period that was merely 18 hours long, immediately before landfall. Following landfall, societal impacts were moderated by the fact that the storm hit a low-population region of Mexico.

According to the National Hurricane Center, John made landfall around 9:15 p.m CST in Mexico on Monday evening (Sept. 23, 2024), which was 0315 UTC on Tuesday morning, Sept. 24, in universal time.

On Tuesday, Sept. 23, Hurricane John was forecast to dissipate over land, and for part of Tuesday, the National Hurricane Center stopped issuing advisories for it. However, the storm’s remnants moved over the Pacific Ocean again where the system regained hurricane strength (65-knot winds) by Thursday, Sept. 26. The next day, John weakened to tropical-storm strength before striking Mexico a second time at a location west of its first landfall earlier in the week.

For more information about the IMERG algorithm and the NASA Global Precipitation Measurement (GPM) mission that IMERG is part of, please visit the GPM website: https://gpm.nasa.gov. Credits: Image and caption by Owen Kelley.

References:

NOAA Historical Hurricane Tracks webpage, https://coast.noaa.gov/hurricanes.

NHC, Discussion #3, https://www.nhc.noaa.gov/archive/2024/ep10/ep102024.discus.003.shtml. Intensity was 40 kt at 3AM CST on Monday, Sept. 23, 2024. Intensity was forecast to increase 20 kt in the following 24 hours.

NHC, Discussion #7, https://www.nhc.noaa.gov/archive/2024/ep10/ep102024.discus.007.shtml. Intensity increased to 105 kt, making John category 3 as of 9PM CST Monday, Sept. 23. The intensity estimate was influence by passive-microwave satellite data (SSMIS) and by T-values from infrared geosynchronous satellite data.

NHC, Discussion #8, https://www.nhc.noaa.gov/archive/2024/ep10/ep102024.discus.008.shtml. Landfall time was 9:15PM CST Monday, Sept. 23, with 105-kt category-3 winds. This time is the same as 0315 UTC, Tuesday, Sept. 24, 2024.

NHC, rain potential images, posted on https://www.nhc.noaa.gov. These images were issued between Sept. 23, 2024, 1200 UTC and Sept. 25, 2024, 1530 UTC. These images showed a narrow area with 20-30 inches of 3-day rainfall accumulation along the coast of the Mexican states of Oaxaca or Guerrero.

• *The programs and services offered by George Mason University are open to all who seek them. George Mason does not discriminate on the basis of race, color, religion, ethnic national origin (including shared ancestry and/or ethnic characteristics), sex, disability, military status (including veteran status), sexual orientation, gender identity, gender expression, age, marital status, pregnancy status, genetic information, or any other characteristic protected by law. After an initial review of its policies and practices, the university affirms its commitment to meet all federal mandates as articulated in federal law, as well as recent executive orders and federal agency directives.

Since 2014, NASA has run the IMERG multi-satellite algorithm to estimate how much precipitation falls over the globe every 30 minutes on a grid with 0.1-degree resolution in latitude and longitude. Each grid box is approximately a square seven miles to a side at the Equator.

Every few years, the team of scientists that develop the IMERG algorithm deliver a new version of the software. They give the algorithm code to the NASA Goddard Precipitation Processing System (PPS) to run. Analysts and computer programmers with George Mason University’s CEOSR assist with running the algorithm and detecting anomalies in its output. When a new version of IMERG is delivered, NASA goes back over 23 years of satellite observations to give updated precipitation estimates that cover every 30 minutes of that entire period.

In 2024, NASA released Version 7 of the IMERG algorithm. An example accumulation from this version of the algorithm is shown above, covering five and a half days of observations of Hurricane Debby. Debby impacted the United States East Coast in August 2024. For part of that time, Debby was stalled off the coast of South Carolina, which means that the hurricane brought on shore a great deal of moisture which fell as large rainfall accumulations.

CEOSR team members used ancillary datasets to help evaluate the quality of the IMERG precipitation accumulations of Hurricane Debbie. The red triangles are locations of preliminary tornado reports associated with Hurricane Debbie according to the NOAA Storm Prediction Center. The violet-colored bow ties are locations where the USGS measured major
flooding at a steam gauge, as reported by the NOAA Water Prediction System. The orange and red squares are locations where rain gauges measured accumulation in excess of 14 or 17 inches, respectively. These gauge data were published in a discussion by the NOAA Weather Prediction Center.

The image’s filled contours in the image show precipitation estimates from the near real-time run of Version 7 IMERG. These near real-time estimates are known as the Late IMERG run. Weather forecasting agencies, disaster management organizations, and researchers can download these estimates approximately 16 hours after the satellite observations were collected. The path taken by the low-pressure center of Hurricane Debby is show with the white dashed line as reported in real-time by the National Hurricane Center.

As can be seen in the image, the maximum accumulation reported by the rain gauges occurred mostly over South Carolina within 50 miles of the coast. Many of the tornadoes and flooding stream gauges were also located this close to the sea. In contrast, the near real-time accumulations from IMERG were largest further inland or off the coast.

Because IMERG relies heavily on satellite estimates, it does not always capture the fine-scale structure of precipitation from severe storms with sharp precipitation gradients. Over much of the world’s oceans and even parts of continents, ground-based precipitation estimates are sparse, so there is interest in improving the IMERG algorithm based on comparisons with ground data where plentiful, such as over much of the United States.

CEOSR researcher Owen Kelley examined Version 7 IMERG accumulations from other hurricane impacting the United States coast in prior years to see if the new Version 7 algorithm always has errors matching the spatial pattern seen here with Hurricane Debby (2024). He determined that the new algorithm does not habitually make this mistake. NASA and CEOSR researchers will continue to evaluate the new version of IMERG with the aim of helping NASA to release an improved version in approximately 2 years.

• *The programs and services offered by George Mason University are open to all who seek them. George Mason does not discriminate on the basis of race, color, religion, ethnic national origin (including shared ancestry and/or ethnic characteristics), sex, disability, military status (including veteran status), sexual orientation, gender identity, gender expression, age, marital status, pregnancy status, genetic information, or any other characteristic protected by law. After an initial review of its policies and practices, the university affirms its commitment to meet all federal mandates as articulated in federal law, as well as recent executive orders and federal agency directives.