Extreme General Storm Precipitation

Late Spring Snow Storm Hammers Parts of Colorado

By | Extreme General Storm Precipitation | No Comments

The past couple of days in Colorado have been quite dramatic in terms of the weather, with temperatures in the mid 70s to low 80s earlier in the week replaced by temperatures in the low 40s, well below normal for this time of year. The dynamics of the atmosphere made for some stark differences across the United States. Above normal temperatures were present across most of the Eastern U.S. with widespread heavy rainfall and severe thunderstorms from the Southern Plains through the Upper Midwest and much cooler temperatures and snow for the western U.S.

Figure 1: A home in Estes Park, Colorado, covered in 25 inches of snow Friday. (NOAA ESRL)

A strong upper level low moved into western Colorado on Thursday morning from the northwest. It initially brought snow for higher elevations and rain for lower elevations. As the day went on, snow began to accumulate in the mountains of Colorado and Wyoming and rain became mixed with snow in areas of the front range. A frontal boundary situated from the Southern Rockies to the Great Lakes region produced widespread severe storms, prompting the SPC to issue an enhanced to high risk thunderstorm outlook that included south-central Kansas and northwestern Oklahoma. There were even a handful of severe storm reports in southeastern Colorado, including a tornado. Several reports of severe thunderstorms, tornadoes, and flash flooding were reported Thursday afternoon and evening across the Southern and Central Plains.

Figure 2: Convective Outlook issued by the Storm Prediction Center for Thursday 5/18/17.

Image Credit:

The meteorological setup across the U.S. is shown in Figure 3. This was a cold, slow moving system, perfect for large amounts of snowfall in Colorado. The heaviest amounts were reported in north and northwest portions of the state, ranging from feet in the mountains to several inches in the plains of northeastern CO. Winter storm warnings were issued late Wednesday afternoon and again early Thursday morning for central and north central Colorado as rain began to turn to snow. The official snowfall total reported at Colorado State University in Fort Collins was 5.8”, which is a daily record for May 18th (old record just a trace in 1915 and 1960), and the highest daily total for any day after May 10th. The daily precipitation of 2.77″ is a daily record for May 18th (old record 1.83″ from 1915), more than average during the entire month of May, and the 3rd largest single May day precipitation on record.

Figure 3: Forecast for Thursday 5/18/17 issued by the National Weather Service

Image Credit:

Several areas experienced large amounts of snowfall, while other areas only received a trace. Largest amounts were in Ward, CO with an astounding 41.7 inches of snow; however, Denver officially received only a trace. Areas that experienced heavy snowfall were heavily impacted. Tree damage was widespread, as leafed-out branches broke under under the weight of the very dense, sticky snow. Schools were closed, roads became dangerous due to slick conditions, and several graduations were postponed as a result of this storm.

Figure 4: Accumulation amounts for Northeastern Colorado over a 72-hour period.

Figure 4 shows accumulation amounts for Northeastern Colorado. Because this was a snow event, the amounts reflected in Figure 4 are water equivalent amounts with a maximum amount of 3.47 inches and largest amounts east of the I-25 corridor between Denver and Greeley. This snow storm brought wet, heavy snow and the amounts reflected in Figure 4 indicate a substantial water accumulation in a brief period of time.

Spring snowstorms are not unusual in Colorado, but this is remarkably late in the year. The unusually cold temperatures are underscored by a freeze warning issued by the National Weather Service for Friday night into Saturday morning for Central Colorado, with impacts to crops and vegetation likely. The system has already begun winding down as it continues to move out of the area, but residents in the Southern Plains should once again brace for another day of severe weather. For more on extreme precipitation events across the U.S., please continue to monitor this space from MetStat.

Ozarks Hammered with Second Major Flood in Last 18 Months

By | Extreme General Storm Precipitation, Extreme Local Storm Precipitation, MetStorm | No Comments

This past weekend, the Ozark Mountains of southern Missouri and northern Arkansas experienced two days of near-continuous heavy thunderstorms. The sustained heavy rain over the steep hills of this region led to rapid runoff and a dramatic rise in local creeks and streams. Numerous towns and cities from central Arkansas to southern Indiana have been inundated over the past few days by floodwaters that, in some cases, have reached record heights. While floodwaters have begun to gradually recede in many locations, another ongoing (but fortunately less extreme) rain event will keep rivers elevated for several more days.

Figure 1. The town of Eureka, MO, a suburb of St. Louis, under record floodwaters from the nearby Meramec River. The river crested at an all-time high of 46.11′ on 2 May 2017.

As is the case with any extreme flooding event, numerous factors contributed to the disaster, but meteorological conditions are foremost. Late last week (on and around April 28, 2017), a sprawling low pressure system spun into life along the southern Rocky Mountains and drifted northeastward. This powerful system inspired a huge variety of spring weather across the middle portion of the United States, with severe and tornadic storms across the southern Plains, a record-strong late-season blizzard in the High Plains of Kansas and Colorado, and heavy rain across much of the lower Missouri and Mississippi River valleys (Figure 3). Below, a spectacular GIF of infrared satellite images from the new GOES-16 shows thunderstorms blossoming across the central part of the country before the main upper level low pushes through. (See also this YouTube video of the lightning from these storms, as captured by GOES-16.)

Figure 2. Thunderstorms blossom from the Ohio River Valley to Texas during the overnight hours of April 28-29, 2017. (Image credit:

Figure 3. Storm-total precipitation across the Ozarks region from the morning of April 28 to the morning of May 1. A large band of 6″+ of rain occurred over hilly terrain, leading to substantial flooding across the region.

While the tornadoes proved deadly and the blizzard may have damaged some sensitive vegetation, by far the most devastation from this storm was associated with the flooding. Figure 4 shows a snapshot of radar across the region from the evening of April 29, along with a selection of storm reports from the April 28-May 1 period. All kinds of severe weather occurred over this period, but note that flooding was widespread from MO/AR through the Ohio River Valley.

Figure 4. A snapshot of radar from the evening of April 29th and an assortment of local storm reports across the Ozarks and Ohio River valley. Image credit:

The most intense flooding was concentrated across the Ozarks of southern Missouri, northern Arkansas, and far northeastern Oklahoma, where 14 river gauges hit all-time record crests this week. Of course, factors such as floodwall and levee development along rivers can certainly affect the flood depths achieved — the more a river’s floodplains are restricted, the higher the water must go. However, we at MetStat like to examine the precipitation factor from a frequency perspective to determine just how extreme rainfall events are (Figure 5).

Figure 5. The Average Recurrence Interval (ARI) for the 72-hour rainfall from April 28 to May 1, 2017.

As the Average Recurrence Interval (ARI) analysis above indicates, some areas experienced rain on par with a 1-in-100 year event or more, particularly across the Ozarks and on into southern Illinois. This product compares the observed rainfall to rainfall frequency distributions in order to illustrate the relatively rarity of an event. On average, a 100-year ARI event would be expected to occur just once every 100 years for a given point. This event brought an impressively large area of 100- to 1000-year rains (reds to purples in Figure 5), which certainly played a leading role in producing such catastrophic flooding.

To some, this storm may seem like a case of déjà vu. A remarkably similar storm system brought rains of remarkably similar magnitude to this same region in late December 2015. As the comparisons below show, the rains in 2015 were slightly heavier, more widespread, and further northwest than the ones this time around:

Figure 6. Comparison between the 72-hour storm total precipitation from the December 26-29, 2015 event (left) and the April 28-May 1, 2017 event (right).

Figure 7. Comparison between the Average Recurrence Interval product for the storm of December 26-29, 2015 (left) and the storm of April 28-May 1, 2017 (right).

Clearly, while the 2015 heavy rain was more widespread, the 2017 event had more intensely concentrated pockets of heavy rain over hillier terrain, and this likely contributed to the extreme nature of recent floods. While these two very rare events occurred within a short time of each other, it must be kept in mind that recurrence intervals just represent a probability, and that large events can happen at any time. This past weekend’s event is the first major event of the season, and more are sure to follow. We are constantly monitoring extreme precipitation events all around the country, so monitor this space for updates from MetStat.

Heavy Precipitation in the Northern Sierras Leads to Oroville Dam Crisis

By | Extreme General Storm Precipitation | No Comments

A dominant story in news headlines over the past few days has been the crisis at California’s Lake Oroville dam, the tallest earthen dam in the United States. Buoyed by weeks of heavy precipitation throughout much of central and northern California, the reservoir for the first time in its 49-year history exceeded its capacity and began to overflow into the emergency spillway (Figures 1 & 2). The temporary evacuation of some 200,000 people occurred on Sunday when engineers became alarmed by the rate of erosion at the lip of the emergency spillway. This came just days after standard releases from the dam began to cause serious erosion of the main spillway, necessitating the reduction of flows while engineers inspected the damage. While the erosion of the main spillway has apparently stabilized as of this writing, the emergency spillway remains potentially unstable if needed again before proper repairs can be made.

Figure 1. An aerial view of the severely eroded main spillway of Lake Oroville on Saturday, 11 Feb 2017. Water can also be seen running over the emergency spillway above and to the left of the main spillway. Image credit: Wikimedia Commons.

Figure 2. Lake Oroville water storage levels in 2016-2017 as compared to other historically wet and dry years. Image source: CA Dept. of Water Resources.

As indicated by the water storage graph in Figure 2, water levels have spiked sharply multiple times already this water year (water years run from October to September). These spikes are the result of a series of powerful low pressure systems that have produced copious amounts of precipitation over the Lake Oroville catchment area, particularly in the mountains. Since mid-December, there have been at least four major storm systems that produced widespread storm-total liquid precipitation totals in excess of 10″ (Figure 3). Above 6,000-8,000 feet in elevation, much of this moisture remains locked up as snow and will only gradually contribute to the flow of water into Lake Oroville. However, as evidenced by the time series below, taken at a point at just 3,300′ elevation, the extremely heavy precipitation has also occurred at lower, warmer altitudes, where almost all of it becomes immediate runoff.

Figure 3. A time series of analyzed daily precipitation values at a point in the mountains upstream of Lake Oroville. Image source: PRISM Explorer.

Based on this time series, MetStat decided to analyze four recent weeklong periods of precipitation in the Oroville drainage area: 10-16 December 2016, 5-11 January, 19-25 January, and 4-10 February. Gridded precipitation data was acquired from NOAA’s Multi-sensor Precipitation Estimates (MPE) and compared to precipitation frequency grids from NOAA’s Atlas 14 Vol. 6 to produce the Average Recurrence Interval (ARI) product. The ARI shows the approximate rarity of an event expressed in terms of the average amount of time that would be expected to pass between events of identical magnitude. In other words, an ARI value of 200 years would indicate a storm that would only occur once every 200 years on average. The images below show the storm total precipitation for each event on the left and the 7-day ARI on the right.

Note that while no individual storm is incredibly rare individually, the combined effects of a 1-2 year storm (storm 1), a 10-25 year storm (storm 2), a common but heavy storm (storm 3), and another 10-25 year storm (storm 4) all in succession has led to some truly staggering rainfall totals. This is illustrated most starkly in the 45-day precipitation totals (from 28 Dec to 10 Feb) and 45-day ARI analysis, which conveys the rarity of seeing this much precipitation in the course of a month and a half (Figure 5).

Figure 5. 45-day total precipitation (left) and and 45-day average recurrence interval (right) for the Lake Oroville drainage area.

The precipitation during this period was definitely impressive: over 20″ of liquid has fallen on all of the hilly and mountainous terrain in the region, and a substantial area upwards of 60″ of liquid is analyzed in the heart of the mountains! Such huge precipitation values are consistent with average return intervals of anywhere from 10 to 250 years in the Oroville region, underscoring the infrequent nature of such a persistently wet pattern in the Northern Sierras. It is worth reiterating that much of that high-elevation precipitation remains in the form of snow, but this highlights the issue at hand: Lake Oroville will be dealing with enormous amounts of water for months to come. In addition to gradual snowmelt over the next several months, the wet pattern looks to continue for the immediate future. The Weather Prediction Center is forecasting another 4-10″ of precipitation to fall over much of the basin in the 7 days to come (Figure 6; see also the WPC website).

Figure 6. A Google Earth view of the WPC’s 7-day precipitation forecast, valid from 15-21 Feb 2017.

With a near-record snowpack building in the Northern Sierra Mountains and another 6-12 weeks of California’s wet season still to come, the crisis at the Lake Oroville dam will likely be ongoing for weeks. Officials are confident that the spillways can be reinforced enough to survive the season, but another extended stormy period could still be a burden and potential threat to water infrastructure all throughout the Sierra Nevada. This event emphasizes the importance of the hydrometeorological expertise that MetStat provides to help design, engineer, and operate safe water infrastructure.

Data Sources:
HDSC Precipitation Frequency Grids
AHPS Precipitation Analysis

Aerial view of spillways
Lake Oroville storage
PRISM precipitation time series

Coastal Washington and Oregon Deluge Sets Up Region for Wettest October on Record

By | Extreme General Storm Precipitation | No Comments

Earlier this month (October 2016), parts of the coastal Pacific Northwest faced nearly a week of heavy rains and high winds as a series of storms crashed ashore. An intense Pacific jet stream, amplified by interaction with the remnants of Typhoon Songda, carried copious amounts of moisture into the coastal mountains of Washington and Oregon, with the most intense rainfall occurring October 13th-17th. Forecasts of widespread hurricane-force wind gusts failed to materialize, but localized wind gusts were nonetheless impressive: peak wind gusts were 70 mph near Honey Lake, CA, 89 mph near Incline Village, NV, 94 mph near Megler, WA, and 103 mph near Oceanside, OR. In addition, two rare tornadoes managed to touch down in the coastal Oregon cities of Manzanita and Oceanside, with substantial structure and tree damage reported in Manzanita.


Noteworthy though the winds may have been, the widespread heavy rainfall was the most impactful aspect of these storms. Our preliminary MetStorm analysis of regional rainfall shows just how widespread the heavy rainfall was, particularly in mountainous areas. Many high-elevation locations recorded in excess of 6″ of rain during this 4-day period, with our analysis indicating a local maximum of up to 20″ in the Olympic Mountains of northwestern Washington.


Even in lower-lying areas between mountain ranges, including such cities as Portland, OR and Seattle, WA, rainfall was substantial enough to push several areas towards their wettest Octobers on record. The rainy season is usually just ramping up through October in the Northwest, so several-inch lowland rainfall events in October are not particularly common, although they wouldn’t be out of place later in the season. The majority of the area received rainfall consistent with what one would expect to occur at least once a year. However, isolated pockets in the mountains and in the Puget Sound region saw much rarer heavy rains, consistent with recurrence intervals in the 10- to 1000-year range. This is particularly impressive when considering that the heaviest rains in these areas generally tend to occur between November and April.


MetStorm’s mass curve analysis illustrates how this multi-day event unfolded. The point of heaviest rainfall in our analysis is isolated, and two fields from that location are plotted below: hourly rainfall values throughout the event (shaded), and cumulative rainfall since the beginning of the event. Several distinct low pressure systems, including one associated with the remnants of what had been Super Typhoon Songda in the west Pacific Ocean several days earlier, brought distinct waves of rainfall to the region. While hourly values never exceeded 1″, the persistence of moderate to heavy rain over such a lengthy period added up–to over 20″ in this case.


Flooding fortunately proved to be minimal with this event, but many areas in the Northwest are now dealing with very saturated ground heading into what forecasters expect could be a wetter-than-average winter. If rainfall events of this magnitude continue to occur this winter, more severe flooding and mudslides are certainly a possibility.

Please note that the maps presented here are preliminary and will be updated when new data become available. If you are interested in this product, or any other product from our MetStorm Precipitation Analysis tool, please email us or send us a message though our contacts page here.

-MetStat Team

From Hurricanes to Thunderstorms: Louisiana’s Storms in Perspective.

By | Extreme General Storm Precipitation, MetStorm, Uncategorized | No Comments

Louisiana is no stranger to natural disasters. From droughts to flooding, since the year 2000 Louisiana has endured 8 natural disasters with over $200 million in estimated economic impact, and 22 FEMA major disaster declarations. To no surprise, the most famous, costliest, and deadliest natural disaster to hit Louisiana was 2005’s Hurricane Katrina. The most expensive hurricane to ever hit the United States, Katrina resulted in 1,833 deaths and economic impacts around $150 billion. At the time of landfall, Hurricane Katrina was a category 3 hurricane with sustained hurricane-force winds extending 120 miles out from its center. Aided by heavy wind and rain, the massive storm surge created by Katrina breached multiple levees and left large swaths of New Orleans underwater within moments of the initial breaches. With that in mind, it is understandable just how extensive the sheer amount of damage was.

Given these impressive stats, it’s hard to imagine another non-hurricane natural disaster that could even come close to having such an impact within the state of Louisiana. Earlier this month, a much less exciting weather phenomena – in the form of a broad area of low pressure – settled in over the American Southeast. This allowed consistent thunderstorm development from August 11th to the 14th. This slow tide of steady rainfall dropped well over 20 inches of rain throughout Louisiana (compare this to the roughly 10″ totals from Katrina) and ultimately lead 13 deaths and has left tens of thousands homeless. In what has been called the worst disaster since Hurricane Sandy, the onslaught of thunderstorm rainfall created flooding virtually unheard of even within a state that by some measures is the wettest state in the country.

Aftermath of Hurricane Katrina (left), compared with recent flooding near Baton Rouge (right).

How did a low pressure system, spinning up thunderstorms that are seemingly mere ordinary afternoon storms, dump 2 to 3 times as much rainfall as Katrina in only a matter of days? The answer lies in its organization. A mesoscale convective system, or MCS, is an intricate structure of thunderstorms that allows each individual storm to become part of a system larger than itself. This organization can take many forms, and usually means that the system as a whole is large and long-lived. Check out the radar reflectively loop over the southern Mississippi Valley in the days of heavy rainfall:


This organization of storms rotated around itself and continually dropped rainfall in both the Baton Rouge and Lafayette areas. Below is a plot generated by MetStorm showing the total rainfall over the 96-hour lifespan of the MCS.


A mass curve plot was also generated by MetStorm, for the area that received the highest total rainfall within the analysis time. Note that for multiple hours across the first two days of the storm event rainfall values exceeded 1″ per hour, and that even after the largest storms had passed, the area still received steady rain for almost another 48 hours.


Of course, flooding is not only apparent in precipitation data, but in river gauges as well. The first plot below shows river height in feet of the Mississippi River near Baton Rouge. Over the course of about 24 hours, the Mississippi rose roughly five feet. For comparison, levee breaches and heavy rain during Katrina rose the Mississippi river at New Orleans by nearly 16 feet in less than 12 hours. Smaller rivers, like the Comite River also plotted below, were subject to the largest increases in river height. In the same 24 hour span in which the Mississippi increased, the Comite River rose from just a foot or two to over 25 feet, shattering the previous gauge height record set in 1961 by over a foot. The large increase in river heights also correspond to the hours with the largest amounts of precipitation, seen in the mass curve plot above.


Finally, in assessing the rareness of this flooding event, we calculated the average recurrence interval of the maximum amount of rainfall at each grid point for both 1- and 24-hours. Diagnosing the maximum ARI value over a 1-hour time span reveals a maximum grid cell value of 94.81 years (i.e. the one hour rainfall maximum has a ~1 in 95 chance of occurring in a given year). While rare, this value is not exceptional in terms of causing such an extreme flooding event in Lousiana. However, paired with the 24-hour ARI analysis, you’ll notice the number of areas that had 24-hour rainfall totals so high that they would only be expected less than once every 1000 years. A single thunderstorm (usually producing rainfall in a fixed location for less than an hour) did not make this event what it was, but rather the large MCS that organized thunderstorms to produce lasting, steady rainfall for days on end in the same locations in the state of Louisiana.



The aftermath of the Louisiana floods have given pause to many residents in the state, and determining how to rebuild after another major flood will be a difficult challenge. In the weeks to come, there will likely be a lot of tropical storm activity in the Atlantic, and we hope that Lousiana is spared from any major tropical storm that finds its way into the Gulf Coast.

Please note that the maps presented here are preliminary and will be updated when new data become available.  If you are interested in this product, or any other product from our MetStorm® Precipitation Analysis tool, please contact us at or through our contacts page at here.

-MetStat Team

40th Anniversary of the Big Thompson Flood

By | Extreme General Storm Precipitation, Extreme Local Storm Precipitation, MetStorm | No Comments

An afternoon thunderstorm situated in just the right place can spark a chain of events that can completely change a community and how it learns to respond to a flooding disaster. Such was the case the evening of July 31st above the Big Thompson Canyon in Colorado, 60 miles northwest of Denver. The thunderstorms, that ultimately killed 145 people and resulted in $40 million of damages, dumped over a years-worth of precipitation in a very short amount of time.


Many people are able to recall this horrific event, due in no small part to luck and quick-thinking. Outrunning a wall of water is an impossible feat in a river canyon, and the unfortunate truth to the Big Thompson flood is that those who attempted this mostly perished. Climbing the canyon walls to safety, however, gives one a much greater chance of survival.


On the evening before Colorado celebrated its 100th birthday, a mass of storms began to set up and take root right above the Big Thompson Canyon. This relatively stationary storm system began its downpour directly over the river, and within a matter of hours completely changed the the surrounding landscape. Below is a mass curve plot, generated by MetStorm, that displays the incremental and accumulated precipitation in the area of heaviest estimated precipitation (a total of 15.6 inches of rainfall over a two-day period).


West of Loveland, throughout the canyon, storm rainfall totals above 10″ stretched from Glen Haven to the north to the border or Rocky Mountain National Park to the south:


This incredible amount of rainfall in such a short amount of time is undoubtedly a rare event. To assess just how rare this amount of rainfall was, an analysis of the average recurrence interval, or ARI, of the storm was performed by the MetStat team. Below, the MetStorm-generated ARI map for the 3-hour period of maximum rainfall for each point on the map shows that for much of the area in and around the Big Thompson Canyon, the amount of rainfall that pummeled the canyon has a less than one in one thousand chance of occurring in any given year.


The Big Thompson flood remains Colorado’s deadliest and one of its most costly. The lessons learned from this event still resonate with Coloradans; and the communities within now know how to respond to such a disaster.

Please note that the maps presented here are preliminary and will be updated when new data become available.  If you are interested in this product, or any other product from our MetStorm® Precipitation Analysis tool, please contact us at or through our contacts page at here.

-MetStat Team

Stunning Microburst in Phoenix

By | Extreme General Storm Precipitation, Extreme Local Storm Precipitation, MetStorm, Uncategorized | No Comments

Phoenix, along with much of the rest of the country, has been battling with excessive heat for most of the summer. In the southwest this dry heat, combined with their summer monsoonal rainfalls, can create a virulent effect in the atmosphere that accompanies the rain, known as a microburst.

Photo Credit: Bruce Haffner

Microbursts form as rain from thunderstorms enter hot, dry air underneath them. This air causes raindrops to evaporate, and in the same manner as hanging around after taking a dip in a pool can make you shiver, it cools the surrounding air. Already being relatively colder to begin with, this cooling by evaporation (or: evaporative cooling) makes the downdraft of rainfall under the storm accelerate. This is because cold air is denser than hot air, causing it to cascade towards the ground faster and faster as more rainfall evaporates. The picture above illustrates this effect perfectly underneath a large thunderstorm producing very heavy rainfall. Once this rush of precipitation and cold air hits the ground, it has nowhere else to go but out horizontally, which is also noticeable in this picture. The violent outflowing air can kick up dust and debris along the way, creating another weather phenomenon called a haboob.

A quick MetStorm analysis on the thunderstorm that produced this incredible display shows 1.42 inches of rainfall falling between 5 and 6pm the evening of the 18th. This is not an uncommon occurrence near Phoenix during the monsoon season. The average recurrence interval for 1.42 inches of precipitation falling in a one hour timespan in this location has a probability of occurring once in ten years.


This storm is an excellent demonstration of natures fury when all the right ingredients come together and produce a visually stunning phenomena.

Please note that the maps presented here are preliminary and will be updated when new data become available.  If you are interested in this product, or any other product from our MetStorm™ Precipitation Analysis tool, please contact us at or through our contacts page at here.

-MetStat Team

Deadly West Virginia Floods

By | Extreme General Storm Precipitation, MetStorm | No Comments

The last weekend in June, a series of large thunderstorms produced historically heavy rainfall across much of West Virginia, ultimately resulting in the loss of 23 lives. The mountainous, complex landscape of West Virginia makes flooding an especially dangerous scenario, because rain water rushes down narrow and steep valleys, within a matter of moments generating a swell of water that can wash away vehicles and homes before people have time to react. The heaviest rains over this weekend mainly effected the counties that contained this type of rugged terrain.


Lewisburg, located in Greenbrier County, was one of the towns effected by the heaviest of rainfall. With an average of 3.74″ of rain over the entire month of June, the roughly 9? of rain that the town experienced over a matter of days adds perspective to just how extreme that amount really is. And with an average of 40? of rainfall falling annually, the storms that passed over the county from June 21st-24th produced nearly one quarter of the rainfall Lewisburg would expect in a given year.

Below is NEXRAD radar imagery from the evening of June 20th to the morning of the 24th (84 hours total). Notice that across this more than three-day time span, parts of West Virginia were almost constantly under of some sort of storm rainfall. Also note that the afternoon and evening of the 23rd correspond to when thunderstorms consistently developed in the same area and moved in a similar fashion across the state, continually dumping precipitation across the same counties. Known as training in meteorology, this phenomenon is also present in our MetStorm analysis of Texas/Oklahoma heavy rainfall from a couple of weeks ago. Much like this previous storm analysis, heavy rainfall at night caught many towns off guard, and combined with the swiftness of onset flooding, created disastrous consequences and has lead to a massive recovery effort.


Our MetStorm analysis was run for the three and a half days that roughly correspond to the length of heavy precipitation. Below is the mass curve time series plot showing incremental and accumulated precipitation for the area of heaviest rainfall – located in Greenbrier County – in UTC time. For reference, UTC, or Greenwich Mean Time, is four hours ahead of Eastern Time during the summer, so midnight Eastern Time corresponds to 4 UTC. While there are times of marginal rainfall throughout the time series, by far the most amount of rain fell during the second half of the 23rd. Above we mentioned that this time period saw training thunderstorms repeatedly unleashing rainfall throughout West Virginia. At this site, the largest hourly value of rainfall was about 2 inches in one hour. And again for perspective, Lewisburg and surrounding areas receive on average ~3.74? of rainfall during the entire month of June.


Focusing more on this 24-hour time period of heaviest rainfall, our next MetStorm analytic is of the total amount of rain that fell during this time over the entire area analyzed, which is plotted below. The maximum 24-hour precipitation across our analysis area is 8.48?, again focused near Lewisburg in Greeenbrier County. This area, unfortunately, contains some of the most mountainous terrain in West Virginia.


Expanding on this analysis of 24-hour rainfall totals is a map of the MetStorm-generated Average Recurrence Intervals, or ARIs. ARI is the probability of the occurrence of the total recorded rainfall amount over a specified duration in any given year. Given all that we have discussed in this post, we already expect this storm event to be a rare event, but 24-hour ARI values for a large swath of West Virginia show an event that is exceptionally rare, with many areas experiencing rainfall that has a less than 1 in 1000 chance of occurring during any given year.


Determining a disaster mitigation strategy is hard work when you’re dealing with an event that has such a small chance of occurring, especially when historic rainfall creates the type of flood-swept-burning-house scenario you’d expect to find in a movie.

Please note that the maps presented here are preliminary and will be updated when new data become available.  If you are interested in this product, or any other product from our MetStorm™ Precipitation Analysis tool, please contact us at or through our contacts page at here.

-MetStat Team


Extreme Rainfall in Texas and Oklahoma

By | Extreme General Storm Precipitation, MetStorm, Uncategorized | No Comments

More heavy rain fell in the south plains last weekend, continuing a rather long cycle of flooding and dangerous storms across the southern plains over the past couple of months (take a look at our previous MetStorm analyses for April storms in Texas, as well as this excellent NASA write-up of widespread rainfall in Texas and Oklahoma from late-May to early-June). Radar composite imagery for Texas and Oklahoma over the course of June 12-13th is shown below.

radar images via

These storms were oftentimes slow-moving, especially in Oklahoma, and frequently went through dissipating and re-development stages. South of the Dallas/Fort Worth radar station (you can easily spot this radar station as the black dot in the center of the circular area of radar “clutter” in the north-Texas region near Dallas/Fort Worth), observe the line of thunderstorms that seemingly remain stationary from about 6 to 11 UTC (1 to 6 am central time) the morning of the 13th. This area experienced what is known as “training” in meteorology, in which thunderstorms consistently develop in the same area and then move in a similar direction as they mature and eventually dissipate. Areas underneath training thunderstorms thus see significant amounts of precipitation, often in a relatively short amount of time, compared to nearby areas. The implications of this training event are discussed below.

Storm rainfall totals for both Texas and Oklahoma exceeded 10 inches over the course of these two days, as shown below in the MetStorm Storm Total Precipitation map. The two regions that experienced the most amount of rainfall were the areas over and just east of Lawton, Oklahoma, and south of the Dallas/Fort Worth metropolitan area in northern Texas. On top of the previous south plains storms already mentioned, this large amount of rainfall over a two-day time span spelled disaster for homes and infrastructure, particularly for Lawton, where homes needed to be evacuated and extensive road closures occurred throughout the area. An eight-mile stretch of I-45 south of Dallas was closed Monday morning due to storm waters. Check out the MetStorm map to see that one of the core areas of precipitation in Texas fell directly over I-45: the area under the training thunderstorms mentioned above.


Situations like these are often difficult to forecast and are a complicated entity from a disaster management perspective. Below are MetStorm mass curve plots of incremental and accumulated precipitation plotted for the storm centers in both Oklahoma near Lawton (above) and south of Dallas along I-45 (below). The vast majority of the rainfall near Lawton fell approximately 12 hours before the rainfall in Texas occurred, each storm system producing the most rainfall at roughly 10 UTC, or 5am central time, on their respective days. Overnight and early morning flooding events such as these are quite dangerous, as they usually catch communities at their most vulnerable times, and similar events in the south plains this year have resulted in numerous deaths.



The heavy rainfall is also reflected in river flow and discharge data acquired from the USGS National Water Information System for nearby river basins. As an example, below is a time series of water discharge along the Neches River near Neches, Texas, from mid-May to present. Observe that in the hours overnight from the 12-13th of June the rate of water discharge surged from 2,000 cubic feet per second to about 5,500 cubic feet per second. During this very short span of time, the river height rose nearly 3 feet from 13 to 16 feet. Also note past surges in water discharge in late May/early June associated with other heavy rainfall events in the area.

data via

The final MetStorm product for this storm event is a determination of the relative rareness of a rainfall event such as this. This is accomplished through the calculation of an Average Recurrence Interval, or ARI. Simply, the ARI is the probability of the occurrence of the total recorded rainfall amount over a specified duration in any given year. Here we have plotted 6-hour ARI values over our area of interest. In Oklahoma, near Lawton as well as south of Norman near I-35, 6-hour ARI values exceeded 500-year occurrence. And in Texas south of Dallas along the I-45 corridor, the maximum ARI was over 1000-years. In other words, this stretch of I-45 saw heavy rainfall over a 6-hour duration that was so large that the probability of its occurrence in any given year is only one in one thousand.


Sunnier and drier days are in the forecast for the southern plains for the days to come, as is most of the rest of the continental United States as a large upper-level ridge settles itself in for the long-run. With June halfway over we’re entering the thick of summer, which will likely be a welcome change of pace after what was an unusually eventful spring and early summer.

Please note that the maps presented here are preliminary and will be updated when new data become available. If you are interested in this product, or any other product from our MetStorm Precipitation Analysis tool, please email us or send us a message though our contacts page here.

-MetStat Team

Texarkana Flood April 29th – 30th, 2016

By | Extreme General Storm Precipitation, MetStorm | No Comments

On the heals of several other heavy rainfall events throughout Texas and other areas of the southeast, the region comprising northeast Texas and southwest Arkansas saw a major precipitation event that brought widespread flooding, tornadoes, property damage, and even fatalities from April 29th to the early morning on April 30th. The rainstorms initiated ahead of an area of cyclogenesis in central to northern Texas and tracked towards the northeast as the day progressed. These storms were being fed by warm and very humid air originating off of the Gulf of Mexico, as shown in the NOAA surface analysis plotted below.


With plenty of moisture on-hand and a driving force of instability aloft in the form of a shortwave trough passing through the region, large storms were able to initiate and grow rapidly. This also allowed for consistent storm development throughout the day and into the evening/early morning hours of April 30th.


A MetStorm analysis performed for this event shows periods of very high precipitation during multiple time frames within our analysis window. Two time periods of note are 12 and 14 UTC, in which high radar-estimated hourly rainfall on the order of 3 and 5 inches per hour, respectively. Also plotted below is the radar reflectivity at roughly 12 UTC to illustrate how widespread the thunderstorms became.



A MetStorm analysis is advantageous in deciphering how major this storm event was because of its calculation of the Average Recurrence Interval (ARI), or the expected likelihood of a rainfall event over a bounded region and time frame. Below is the maximum ARI calculated across a 6-hour period during the rainfall event for the Texas-Arkansas-Oklahoma region hardest hit by these storms. The darker the shading, the rarer (more extreme) the amount of rain that fell over that 6-hour window. One area in particular (along the Oklahoma and Arkansas border) experienced an exceptionally rare rainfall event, one that likely occurs once every 500-1000 years, and within that an area with an estimate of over 1000 years. The rainfall in northeast Texas was quite exceptional as well.


Compare these areas of large ARIs to the MetStorm rainfall totals recorded across the area using a combination of rain gauge, radar, and satellite retrieval estimates. Close to 16 inches of rainfall fell in the area along the Oklahoma-Arkansas border for the 48-hour window that our MetStorm analysis took place. And again, in Texas, large areas received close to 10 inches of total rainfall. These heavy precipitation bands led to major flooding for large portions of the analysis region.


Please note that the maps presented here are preliminary and will be updated when new data become available.  If you are interested in this product, or any other product from our MetStorm Precipitation Analysis tool, please email us or send us a message though our contacts page here.

-MetStat Team