The first half of August 2026 has brought repeated heavy rainfall to the state of Indiana, with the area in and around the capital city of Indianapolis experiencing particularly strong flooding. Numerous roads are closed, and a bridge in Carmel, IN, just north of Indianapolis, has been washed out. Multiple fatalities have already been reported, and continued storms on Sunday 16 August are exacerbating an already challenging situation.

A sustained river flooding event builds over an extended period of time, and this event is no exception. We’ll start our analysis by looking at the cumulative rainfall graph for Indianapolis dating back to 31 July 2026 as recorded by the invaluable xmACIS2 site. So far in August 2026, Indianapolis has seen three separate days with at least one inch of rainfall, including a whopping 2.09 inches on the 1st of the month. As a result, Indianapolis has experienced more than twice the normal amount of rainfall up to this point in August.

All of that water has to go somewhere, and so it rushes into the streams and rivers of Central Indiana. This plot shows the time series of water flow in the White River just north of Indianapolis dating back to August 1, roughly contemporaneous with the precipitation graph above. Note the logarithmic scale. It’s easy to see how the initial heavy rainfall on the 1st caused the water levels to rise and the flow rates to increase over the next few days, but an extended period of over a week without rain caused a gradual slowdown in the water flow. However, once the rain returned it came back in a big way and the water levels quickly exceeded the system’s capacity to absorb them. Between the 11th and the 15th flow rates increased by a factor of 70.

Perhaps the most important question is: what’s been causing all this rain? Here’s where our satellite tools can provide valuable insight. The CIMSS MIMIC-TPW2 product helps identify areas of elevated moisture by compositing microwave observations of total precipitable water (TPW) with gridded wind data to depict how the moisture evolves over time. Recall that TPW is the measure of how much precipitation in a column of air could be produced if all of the water vapor in that column condensed into liquid and fell to the ground.

Here’s the TPW plot for the contiguous United States on 1 August at 0000 UTC (31 July at 8:00 PM EDT) showing the state of the atmosphere just ahead of that 2 inch rainfall day. Note the strong moist plume that runs up the lower Mississippi valley and into Illinois and Indiana. These TPW values are nearly tropical in magnitude. (That’s no exaggeration: compare the values of the TPW in the Midwest to the bullseye of Tropical Storm Genevieve around 20 N 130 W).

Similarly high levels of TPW have been present over the past several days. Here’s an animation of the CIMSS MIMIC-TPW2 product for Sunday 16 August, covering the most recent round of storms over Indiana. Again, these storms have quite high TPW values, which are made possible by warm August air with a high capacity for elevated moisture levels, significant moisture advection from the Gulf, and high transpiration from the Midwest’s corn fields.

Let’s take a closer look at the most recent round of storms to hit central Indiana. The GOES-19 (GOES East) true color view shows a band of deep moist convection heading east-southeast from Illinois into most of central Indiana. Numerous overshooting tops are visible which become easier to see as the loop goes on; this is more likely due to the setting sun casting longer shadows than it is any trait of the clouds themselves.

The IR sandwich product can be a quite useful tool for assessing the strength of daytime convection. In the IR sandwich, colors representing the infrared window brightness temperatures are overlaid on top of the high resolution visible channel. Lower (warmer) clouds remain unshaded, but deeper (colder) clouds have colors associated with them. Since the overshooting cloud tops are going to be the coldest parts of the cloud, they’re going to appear as a different color. This makes the clouds appear to have more texture than the true color product shows, making it easier for forecasters to identify where the most vigorous convection is. In essence, this unites some of the best aspects of the visible band (high spatial resolution) with the infrared (quantitative information about cloud temperatures).

Satellites can also help identify where the flooding is taking place. Back in July, the CIMSS Satellite Blog discussed the ABI Flood Mapping package deployed by the CSPP Geo group. In short, by comparing geostationary observations of surface water and comparing those to a baseline level of known bodies of water, it’s possible to determine the extent of surface flooding. By applying this technique to the geostationary ABI instrument, forecasters and hydrologists can track flooding on fine timescales and, perhaps more importantly, be more likely to dodge the clouds that would otherwise inhibit the observations. Here’s an animation that shows how that works in practice. This animation stretches from 1500 to 2000 UTC on 16 August 2026 (11:00 AM to 4:00 PM EDT). Tan represents dry land, gray depicts missing data due to clouds, and blue shows where surface water is expected. The other colors represent various degrees of flooding. As the clouds clear we see flooded areas emerge in central Indiana, especially around the White River which cuts northeast to southwest across Indianapolis and central Indiana.

Both the IR sandwich and the ABI Flood Mapping products are available from SSEC’s RealEarth data viewer.

View only this post Read Less

1-minute Mesoscale Domain Sector GOES-18 (GOES-West) Infrared Window and Visible images (above) showed Tropical Storm Lala as it intensified to become a Category 1 Hurricane at 1900 UTC on 15 August 2026 (GOES-18 images | NHC advisory) south of the Big Island of Hawaii. Lala exhibited a rather ragged eye that remained generally cloud-filled — and while the hurricane did not make an official landfall, its northern eyewall did pass across the far southern tip of the Big Island. An overlay of 1-minute GOES-18 GLM Flash Points revealed that there was no intermittent satellite-detected lightning activity until 2323 UTC on 15 August. In terms of wind, some RAWS sites on the Big Island reported gusts in the 60-77 mph range, and Kona International Airport reported a peak wind of 64 knots (74 mph) — with a peak gust of 140 mph occurring at the Mauna Kea summit. Highest wind gusts at other islands included 85 mph in Maui County and 74 mph in Honolulu County.

The coldest overshooting tops surrounding the eye of Lala exhibited infrared brightness temperatures in the -75 to -80 C range (brighter white pixels embedded within dark black regions) — according to a plot of rawinsonde data from Hilo on the Big Island (below), those temperatures roughly corresponded to altitudes near or just above the Forecast Surface air parcel’s Equilibrium Level (EL).

It is noteworthy that the Precipitable Water (PW) value on the Hilo sounding was 2.73 inches — which appears to be the highest PW value on record for Hilo, just eclipsing the prior record maximum value of 2.72 inches on 24 August 1961 (below).

With an unusually moist atmosphere in place, heavy rainfall amounts in excess of 43 inches occurred on the Big Island. The MIMIC Total Precipitable Water product (below) indicated that Lala was drawing rich tropical moisture northward from the Intertropical Convergence Zone (ITCZ) as the tropical cyclone developed and moved northwestward toward Hawai`i.

A sequence of storm-centered Rainfall Rate products derived using microwave data from the ATMS and AMSU instruments on a variety of polar-orbiting satellites is shown below — access to these and other Low-Earth Orbiting (LEO) satellite Direct Broadcast ground station products is discussed in this blog post.

The ample moisture of Hurricane Lala that was helping to produce the heavy rainfall was also producing snow at the summit of Mauna Kea, as documented by web camera video shortly before 0100 UTC on 16 August:

It's happening. Hurricane Lala is currently producing SNOW on the summit of Mauna Kea pic.twitter.com/kwz1RNm0Yl

— Greg Diamond (@gdimeweather) August 16, 2026

A sequence of rawinsonde data plots from Hilo (below) showed that the freezing level (labeled FZL, in cyan) was at its lowest height (15174 feet) at 0000 UTC on 16 August — which was only 1371 feet (418 meters) above the Mauna Kea summit elevation of 13803 feet. Larger snowflake aggregates falling from clouds above Mauna Kea would likely not have melted entirely before arriving at the summit.

_____________________________________

A 3-day animation of GOES-18 daytime True Color RGB + Nighttime Microphysics RGB images (below) displayed the northwestward trek of Lala from 14-16 August.

It is interesting to note that on 14 August, as Lala was still well southeast of the Big Island, a hazy plume of “vog” (volcanic fog) could be seen southwest of the island — which was emitted by Episode 53 of an ongoing eruption of Kilauea (the hot thermal anomaly of the volcano’s lava flow showed up as a small cluster of dark purple pixels in Nighttime Microphysics RGB imagery).

View only this post Read Less

1-minute Mesoscale Domain Sector GOES-18 (GOES-West) Visible images (above) and Infrared Window images (below) showed the slow organization of cloud features associated with Tropical Storm Lala southeast of Hawai’i on 13 August 2026. Overlays of GOES-18 GLM Flash Points indicated that lightning activity was sparse during that time period. Lala was moving through an environment of very low deep-layer wind shear and traversing fairly warm Sea Surface Temperatures — factors which favored further intensification as Lala approached Hawai’i.

During the preceding nighttime hours, before Potential Tropical Cyclone One-C had intensified to become Tropical Storm Lala, VIIRS Day/Night Band images from NOAA-20 and NOAA-21 (below) displayed the faint signature of cloud features — faint because the Moon had not yet risen at that location, so the only source of illumination was atmospheric airglow.

View only this post Read Less

Eyes around the world were turned toward Greenland, Iceland, and Spain as a total solar eclipse tracked across parts of the Northern Hemisphere yesterday, 12 August 2026. If you haven’t already seen the CIMSS Blog post on yesterday’s total solar eclipse from the GOES perspective, be sure to check it out. However, this was a high-latitude eclipse and it was difficult to see the lunar shadow from the standard geostationary view. By contrast, the polar-orbiting satellites of the Joint Polar Satellite System (JPSS) are able to view the eclipse from a much more direct angle. Here’s a static true color view from NOAA21’s VIIRS at 1720 UTC. Note the expected coloration over the eastern contiguous United States, but as you look further to the north you see the darkening of the ground due to increasingly obscured sunlight.

Even better, the JPSS satellites also feature the VIIRS imager. What better way to view the simultaneous day and night of a solar eclipse than the Day/Night Band? Let’s take a look! These images have been plotted via the excellent polar SLIDER resource from our colleagues and friends at the Cooperative Institute for Research in the Atmosphere. These images are North Pole views, with Alaska at the bottom center and Greenland off to the upper right. You can see much of the contiguous United States in the lower right.

The first hint of a partial eclipse began at 1534 UTC (7:34 AM Alaska time, for context). Therefore, this first image, at 1518 shows the pre-eclipse environment. SLIDER overlays previous swaths on the image, hence the nighttime lights in the United States. However, in general nothing too unusual appears to be going on.

The next swath is at 1603 UTC. The eclipse is only a few minutes old at this point, and there’s not really much that is easily identifiable as being an effect of the dimming sun. Still, we see lots of clouds over the North Pole and some clear seas between Greenland and Canada.

By the time we reach 1659 UTC, however, we see a substantial change in the darkness over eastern Siberia. The Day/Night band relies on moonlight when the sun is not available, and when the moon is not present the views are going to be dark. Geometrically, a solar eclipse can only take place during a new moon. Therefore the nighttime parts of the globe are going to be black, and the part of the Earth beneath the moon’s shadow will also be dark. This swath was right at the time the total eclipse first began.

With the next swath at 1744 UTC, we see that the shadow has moved to far northern Greenland. The eclipse reached it maximum at 1746 UTC, so this view represents the eclipse at effectively its greatest extent. We can see this as a dark hole in upper center-right of the next image, where previously Greenland clearly had snow, clouds, and ice.

The JPSS satellites are sun-synchronous, moving from east to west so that they feature largely similar solar characteristics from one swath to the next. Because of this, there are no further satellite views of the eclipse as it moved onward to Spain. Regardless, JPSS gave us a unique perspective of one of nature’s most thrilling (and predictable!) natural phenomena. Stay tuned to the CIMSS Satellite Blog on 2 August 2027, when we’ll be discussing the next solar eclipse which will be taking place over Spain, Africa, and the Mediterranean.

View only this post Read Less