Over the past few days, a staggering amount of water fell across Southeast Texas and Louisiana. All this water disgorged from one of the greatest meteorological events in U.S. history: an intense Category 4 hurricane named Harvey. This analysis attempts to explain how this extreme storm evolved — from landfall to eventual exit from Texas, with emphasis on the rain generation.

The rain accumulation totals

Our strategy here is to look at the behavior of this weather system day by day, from Aug. 25-30. But first, let’s examine a map of the preliminary rain amounts across the greater Texas region (above), over this multiday period from Aug. 23-30.

These values are all subject to continued verification. The deepest water accumulations encompass Houston, Beaumont, Galveston, Pasadena and Port Arthur.

To describe the history of this storm, I have created two key figures, mosaics that show snapshots of the storm, every day for six days. Both of these figures (shown below) can be enlarged by clicking on them.

The first figure displays a surface weather map, with isobars of pressure and the pattern and intensity of rain (colored regions). I have added some details illustrating the movement of the storm (black arrows) and also the position of frontal boundaries (red heavy line = warm front, alternating blue/red line = stationary front). Surface airflow vectors are also shown by tiny orange arrows.

In the next mosaic, I focus on the radar and satellite structure of the storm. The satellite presentation becomes more important after Day 3, and I present a view called “color-enhanced water vapor.” Shown is the water vapor concentration in the middle and upper atmosphere, using colors that clearly delineate saturated, rainy regions from dry areas of airflow.

Day 1. Friday, Aug. 25: Landfall and onset of heavy rain

As a marginal Category 4 hurricane, Harvey blasted ashore across the Central Texas coast. At the same time, its outer rain bands (called feeders) began raking across the upper Texas coast, including Houston. Rain rates also began to pick up near the landfall point, as Harvey’s double eyewall crossed the shoreline.

Harvey created a counterclockwise flow of warm and humid air off the Gulf of Mexico, forming a front along the coast to the east of the storm’s center. This stationary front separated the warm, Gulf air mass from a slightly cooler and drier air mass over land. This setup would factor significantly in the generation of widespread, heavy rain in the days to come.

Already, several processes were creating heavy rains along the upper Texas coast:

  • Feeder bands on the periphery of the storm, which were popping up along the coast, due to “frictional convergence” i.e. the sudden deceleration of gulf air as it encountering land, forcing the air to rise.
  • Air forced upward along the newly created frontal boundary, a process termed “isentropic upglide.”
  • An unstable air mass originating over the gulf, which enabled air to ascend as strong convective cloud turrets producing torrential rain.
  • Concentrations of convective cells over the same locations within the feeder bands (link to Halverson’s first Harvey story here).
  • The exceptionally slow forward speed of the storm.

Day 2. Saturday, Aug. 26: A persistent corridor of heavy rain sets up

Once 100 miles inland, the storm rapidly weakened (dropping from Category 4 to tropical storm), but it also slowed to a crawl. The storm stalled because the “steering winds” (the deep air current in which the storm was embedded) were extremely sluggish. The winds were weak, in turn, because the westerly jet stream was well north of the storm.

From the two figures above, note that the isobars became spread apart and fewer in number (the signature of weakening) and the eyewall’s intense ring of rain diminished significantly. At the same time, the system developed a striking asymmetry: A prominent, intense rainband that set up over about 80 to 100 miles north and east of the storm’s center, over Houston. There this band would remain parked, for a very long time.

This stout band was rooted in a corridor of unstable, very moist (dew points in mid-upper 70s) air streaming off warm gulf water. Additionally, the cyclonic inflow became strongly convergent east of the center, creating a type of “warm sector” more akin to everyday cyclones of the mid latitudes. Deep convective rain cells continuously erupted within this plume, racing north across the warm front, where they received additional uplift. The parade of cells trained constantly over the greater Houston area, with rain rates in the 2 to 3 inches per hour range, and up to 5 to 6 inches in extreme cases.

More than any other reason, this moisture conveyor dumped prodigious quantities of water on the upper Texas coastline. Understanding how it formed, and why it persisted, are key to unlocking the mechanism behind this weather disaster.

Within the warm sector, a strong jet of moist wind from the south, approaching 90 mph, howled along at 5,000 feet above ground — enabling Tropical Storm Harvey to freely slurp Gulf moisture through a gigantic pipe. And the deeper air mass feeding into this pipe was incredibly moist, with precipitable water values (PWV) — how much rain would fall in a 1-meter-square column, measured from the ground to the top of the atmosphere in the 2.5-to-2.8 inch range. At times, PWV surged in excess of 3 inches: We have at least one weather balloon measurement of 3.26 inches, close to the theoretical maximum for this location and time of year.

Day 3. Sunday, Aug, 27: Harvey becomes lopsided

As the relentless, moist conveyor on the storm’s east side continued, big changes affected the overall structure of the storm. As the satellite image shows, a mass of elevated, dry air began circulating into the west and southwest sections of the storm. This air evaporated much of the rain-bearing cloud, creating a very lopsided and asymmetric system: Dry, stable air with clear skies in the southwest portion, and a continued deluge in the eastern and northeast sectors.

The pressure gradient and winds continued to diminish as the storm remained stalled inland. The intense rain band/warm sector circulation was beginning to slide eastward, expanding over western Louisiana, as rain-laden cells continued to train over Houston.

Later in the day, the storm started to migrate back toward the coast, to the southeast. This would shift some of the remaining rain in the circulation’s core back along the middle Texas coastline, and possibly invigorate the storm once the center reemerged over warm water.

Day 4. Monday, Aug. 28: Harvey transforms and is on the move

Harvey slid over the warm gulf waters as dry air surged into the inner core, effectively sequestering a small patch of saturated air in the very center. Meanwhile, the lopsided system continued to inundate eastern Texas and much of Louisiana with its moisture corridor, drawing in air with a PWV of 2.8 inches and converting it into rain at a rate of 3 inches per hour.

Discussions at the NOAA Weather Prediction Center were referring to specific air currents and processes not typical of tropical weather systems, but rather, mid-latitude cyclones, which are normally found farther north: Dry conveyor, warm conveyor, cool conveyor, warm sector, warm front, deformation zone (some of these features are shown by the colored arrows in the figure above). Structurally, the storm was a hybrid type vortex — possessing attributes of both tropical and nontropical cyclones. Per Hurricane Center discussions, the term “tropical storm” was smartly maintained in all advisories and warnings. Yes, Harvey was still very tropical, and most likely continued to possess a warm core. But in many ways, this system had already begun its gradual transition to a nontropical beast.

Day 5. Tuesday, Aug. 29: A burst of core convection

Reemerging over water, Harvey only minimally reintensified, but its transit was marked by a burst of convection in the very center, as clearly labeled in the satellite image. This caused torrential rains to return to the upper Texas coast. The entire system acquired the classic, comma-cloud shape of an extratropical cyclone, so common throughout the year across North America.

Why didn’t “tropical storm” Harvey bounce back and reintensify to any significant extent? I can offer two reasons:

  • The massive intrusion of dry air into the storm’s core limited widespread deep convection.
  • The storm essentially picked up little additional energy over the Gulf waters, compared to land.

All tropical cyclones derive much of their energy from the flux (transfer) of water vapor off warm water. But bear in mind, much of Southeast Texas, being completely inundated, had become an inland extension of the Gulf of Mexico.

When dry land turns into a shallow sea, this can create the “brown ocean effect,” and even fully saturated soil can do the trick. South Florida is a notorious location for this effect, given its extensive waterlogged surface (numerous lakes and marshes). There are documented cases in the scientific literature where tropical cyclones either maintained intensity, or actually intensified, over land … due to the brown ocean. To me, this process seems like a slam dunk for Harvey.

The tremendous volume of condensed water in those storm clouds delivered sufficient latent heat (the heat of condensation, released into the air when water vapor changes phase to liquid droplets) to maintain this storm at or around 40 knots sustained wind, for days over land. From an energy standpoint, “tropical storm” Harvey, while over land, was still behaving very much like a hurricane.

Day 6, Wednesday, Aug. 30: Crossing over to Louisiana

Harvey was finally on the move, exiting Texas to the northeast, en route to the Tennessee Valley, with likely dissipation over the Ohio Valley by week’s end … but not without tangling further with a separate frontal system over the Mid-Atlantic. Flooding rains spread east and north across the mid-South, and Harvey’s full extratropical transition would continue to unfold. The satellite presentation on Day 6 shows a deeply disconnected system, with the mid- and upper-level vortex completely enshrouded in dry air, and warm sector/warm conveyor/rain corridor displaced hundreds of miles east. Harvey had become quite sprawling and asymmetric.

These are the broad strokes of the story. Many details remain blurry, but no doubt legions of researchers and forecasters will analyze and simulate Harvey in the coming months and even years. Harvey’s peculiar hybrid, lopsided structure — with all its rainmaking machinery displaced so far to one side and a massive sweep of dry air on the other — may never have been observed before (on such a large and persistent scale). The storm was also remarkable in terms of its staying power over land, perhaps the best example so far of the brown ocean effect. I rank this storm right up there with Sandy (2012) in terms of structural oddity, severity of hazards and overall impact.

We invite you to share your impressions of this most peculiar storm.