Stark contrast in Eastern U.S. soil moisture following Hurricane Matthew

Stark contrast in Eastern U.S. soil moisture following Hurricane Matthew

Major Hurricane Matthew left a trail of destruction in its wake from the Caribbean up through the U.S. East Coast.  As Hurricane Matthew tracked northward along a large portion of the U.S. Southeast Coast from Florida to North Carolina, the rainfall impacts worsened.  Figure 1 shows the weekly rainfall spanning 4-11 October, ranging from ~2-8 inches along the Florida East Coast to 10-20 inches in the eastern Carolinas.  Since antecedent soil moisture was highest in the eastern Carolinas (Fig. 2), the extreme rainfall led to the most serious flooding in this area.

fig1

Fig. 1.  Weekly rainfall totals from 4 – 11 October 2016.

fig2

Fig. 2.  Total Column (0-2 m) relative soil moisture prior to Hurricane Matthew’s impact on North and South Carolina, valid at 0000 UTC 7 October 2016.

Referring back to the precipitation totals in Fig. 1, we can see that there was a sharp rainfall gradient on the northwestern edge in the Middle Atlantic region.  Interestingly, this gradient in Hurricane Matthew’s rainfall coincided with a pre-existing transition zone between wet conditions near the Atlantic coast and drought conditions further inland from the Appalachians through New England.  The net result was to accentuate the wet-dry contrast already in place.  The animation in Fig. 3 highlights this contrast nicely by presenting the SPoRT-LIS daily total-column relative soil moisture percentiles from 1-12 October.  The percentiles are based off a 1981-2013 daily soil moisture climatology that SPoRT produced from its ~3-km resolution SPoRT-LIS simulation.  By 9 October, notice the incredible transition from excessively wet soil moisture exceeding the 98th percentile (Carolinas through the southern half of Delaware) to extremely dry soil moisture less than the 5th percentile across Pennsylvania into the Northeast (as well as much of the inland Southeastern U.S.).  In fact, total column soil moisture values are less than the 2nd percentile over a large part of Ohio, Pennsylvania, New York, and the New England states, indicative of the ongoing severe drought there.

fig3_loop

Fig. 3. Daily animation of SPoRT-LIS total column relative soil moisture percentile from 1 to 12 October 2016.

Post Frontal Orographic Snowfall Impacts NM

A slow-moving upper level storm system tracked east across northern NM and southern CO on 14-15 December 2015. A weak tap of subtropical moisture ahead of this system provided light to moderate snowfall mainly along the Continental Divide of western NM and the higher terrain running north-south through central and northern NM. Snow accumulations of 3 to 8 inches were reported ahead of and immediately behind the surface front and the mid level trough passage. A classic westerly,upslope flow event developed behind the upper wave as moist, unstable flow interacted with the north-south oriented higher terrain. Winter weather advisories and winter storm warnings were in effect over much of northern NM for the expectation of storm total snowfall of 8 to 12″ with locally higher amounts. Figure 1 depicts the distribution of advisories and warnings over northern NM on the Albuquerque National Weather Service public page.

Advisory and warning map for the ABQ CWA valid 15 December 2015.

Figure 1. Advisory and warning map for the ABQ County Warning Area valid 15 December 2015.

Poor radar coverage over northern and western NM makes it a challenge for assessing winter precipitation patterns and snowfall rates. Figure 2 shows a radar mosaic valid 1800 UTC 15 December 2015 utilizing an enhanced color curve to identify areas of lighter snowfall. Automated surface observations are sparse in this area however there are at least a few observations reporting snow where nothing is present in the radar reflectivity. Webcams at ski resorts serve as an excellent near real-time proxy for visualizing active snow accumulations in these poor radar coverage regions. Additionally, once daily snow accumulation reports from ski resorts aid the verification process following the winter event.

Figure 2. Winter radar mosaic from KABX valid 1800 UTC 15 December 2015. Note the orange circle depicting a large area of poor radar coverage.

The integration of satellite data allows forecasters to supplement these data void areas. The most recent interation of the NESDIS snowfall rate products available at WFO Albuquerque illustrate the snowfall rate derived from radar (Figure 3a) and the snowfall rate available from merging the POES satellite data with the radar data (Figure 3b). Note the grey areas overlaid on the map in Figure 3a indicate areas of reliable radar coverage. The snowfall rate derived from satellite data in Figure 3b clearly shows coverage outside of the area with reliable radar coverage. A very cold and unstable airmass in association with this precipitation suggested snowfall rates in the higher terrain would average between 20-30:1. The 18:1 image in the lower right of Figure 3b indicated rates around 0.4/hr.

FIgure 3a. Radar derived snowfall rate product over northern NM valid 1750 UTC 15 December 2015.

Figure 3a. Radar derived snowfall rate product over northern NM valid 1750 UTC 15 December 2015. Note the grey areas overlaid on the map indicating where reliable radar coverage exists. Upper left (liquid equivalent), upper right (10:1), lower right (18:1), lower left (36:1).

NESDIS snowfall rate product filling in the radar gaps over northern NM valid 1750 UTC 15 December 2015. Note the circles in the upper left image are the location of the webcams in Figure 4.

Although there is sparse coverage of automated surface observations around the higher terrain, webcams from ski resorts can verify the existence of moderate to heavy snowfall. Visibilities in the webcams below suggest snowfall rates higher than those depicted in the NESDIS products – visually, rates look closer to perhaps 1″/hr in the upper right and lower right images (Figure 4). One of our goals of this assessment is to combine information from the webcams with the more quantitative snowfall rate product to better estimate snowfall in data void areas. Snowfall reports from the Chama Railyard indicated 8.5″, Taos Ski Valley 6″, Ski Santa Fe 12″, and Pajarito Mountain 10″.

 

Figure 4. Webcams from across northern NM. Top left (Chama Railyard, yellow circle), Top right (Taos Ski Valley, white circle), Bottom right (Ski Santa Fe, red circle), Bottom left (Pajarito Mt, orange circle).

 

From drought to flooding in less than a week over the Carolinas, as depicted by SPoRT-LIS

A closed upper low over the Southeastern U.S. combined with a deep tropical moisture connection with Hurricane Joaquin led to historic rainfall and flooding over North and especially South Carolina over the weekend.  A wide swath of central South Carolina from the coast to Columbia received over 20 inches of rainfall in the past week (Fig. 1), much of it in the last 2-3 days.  Figure 2 shows the NASA Integrated Multi-satellitE Retrievals for GPM (IMERG) 24-h rainfall estimates displayed in AWIPS II compared to the official NWS/River Forecast Center rainfall estimate for the period ending 1200 UTC on 4 October.  This is erasing the prevailing drought in the Carolinas, which still had moderate to severe drought in the most recent U.S. Drought Monitor weekly product valid 29 September (Fig. 3).

SPoRT’s real-time configuration of the NASA Land Information System (SPoRT-LIS) runs the Noah land surface model to generate a “best modeled” soil moisture estimate at ~3-km resolution for enhanced situational awareness and input to local/regional numerical weather prediction models.  The SPoRT-LIS was assessed during summer/fall 2014 by the NOAA/NWS weather forecast offices (WFOs) at Houston, TX, Raleigh, NC, and Huntsville, AL. Following an expansion to a full Continental U.S. domain, the SPoRT-LIS was also evaluated more informally by the NWS WFOs at Tucson, AZ and Albuquerque, NM this past summer.  The primary areas of utility has been in drought monitoring and assessing areal and river flooding potential.  However, the Southwestern U.S. offices applied SPoRT-LIS soil moisture fields to enhance situational awareness for wildfire and blowing dust situations as well. Overall, a majority of the users during these assessments expressed substantial utility of the product due to a lack of other real-time high-resolution soil moisture products, and were confident enough to use the product as part of operational, public forecasts.

This extreme rainfall event in the Carolinas was captured nicely by the real-time SPoRT-LIS, which depicted some of the most dramatic changes in total column soil moisture ever documented by SPoRT collaborators.  Figure 4 compares the SPoRT-LIS 0-2 m total column relative soil moisture (RSM) from 28 September (left panel) and 5 October (right panel), with Figure 5 highlighting the 1-week change in 0-2 m RSM as displayed in AWIPS II.  The RSM represents how the volumetric soil moisture scales between wilting (0%) and saturation (100%) for a given soil composition, where the wilting point indicates that vegetation can no longer extract moisture from the soil and saturation indicates no more infiltration is possible (thus all new precipitation goes to runoff). Previous experience by the Huntsville WFO found that total column RSM values of ~60% and above tend to indicate an enhanced threat for areal and river flooding over northern Alabama; however, these thresholds can vary depending on river basin properties and regional soil composition.

Values of total column RSM typically ranged from ~25-35% on 28 September, prior to the significant rain event.  However, by 5 October, total column RSM increased to well above 65% in most areas of central South Carolina, and parts of southern and far western North Carolina.  The maximum weekly change in 0-2 m RSM (Fig. 5) exceeds 58% in central South Carolina — a value never documented in the recent years of real-time SPoRT-LIS output!  Most areas of SPoRT-LIS 0-2 m RSM exceeding ~60% correspond to areas of active minor to major river flooding across parts of southern Virginia and the Carolinas, as depicted in the USGS/NOAA river gauge network this morning (Fig. 6).

Finally, SPoRT is producing an experimental daily, real-time soil moisture percentile product based on a 33-year LIS-Noah county-by-county soil moisture climatology.  The soil moisture percentile map indicates where the current 0-2 m RSM soil moisture values lie in the present day’s historical soil moisture distribution for every county in the Continental U.S. The percentile product valid at 1200 UTC 27 September and 4 October is shown in Figure 7.  Primarily dry soil moisture values between ~5th-30th percentile are prevalent across the Carolinas on 27 September, corresponding reasonably well to the U.S. Drought Monitor moderate to severe drought areas from Figure 3. However, after the 10-20+ inches of rainfall over the past week, the 4 October soil moisture percentiles completely reversed across the region, with values > 98th percentile occurring in central South Carolina where the most severe flooding is taking place.  SPoRT plans to develop a brief training module on this percentile product prior to dissemination for display in AWIPS II at NWS WFOs, along with the current suite of SPoRT-LIS fields already available in AWIPS II.

Fig. 1. NWS River Forecast Center rainfall analysis for the week ending 1200 UTC 5 October 20125.

Fig. 1. NWS River Forecast Center rainfall analysis for the week ending 1200 UTC 5 October 2015.

Fig. 2. Comparison of NASA Integrated Multi-satellitE Retrievals for GPM (IMERG) rainfall estimate to the NWS/River Forecast Center analysis for the 24-hour period ending 1200 UTC 4 October 2015.

Fig. 2. Comparison of NASA Integrated Multi-satellitE Retrievals for GPM (IMERG) rainfall estimate to the NWS/River Forecast Center analysis for the 24-hour period ending 1200 UTC 4 October 2015.

Fig. 2. U.S. Drought Monitor weekly drought product valid 29 September 2015.

Fig. 3. U.S. Drought Monitor weekly drought product valid 29 September 2015.

Fig. 3. SPoRT-LIS total column (0-2 m) relative soil moisture valid on (left panel) 28 September, and (right panel) 5 October 2015.

Fig. 4. SPoRT-LIS total column (0-2 m) relative soil moisture valid on (left panel) 28 September, and (right panel) 5 October 2015.  Masked white areas represent water or urban pixels.

Fig. 4. One-week change in SPoRT-LIS total column relative soil moisture for the week ending 5 October 2015, as displayed in AWIPS II.

Fig. 5. One-week change in SPoRT-LIS total column relative soil moisture for the week ending 5 October 2015, as displayed in AWIPS II.  Maximum weekly change value > 58% is highlighted by the cursor position.  Masked black areas represent water or urban pixels.

Fig. 6. USGS and NWS River Forecast Center river gauge plot for the morning of 5 October 2015. River gauges experiencing flooding are indicated by the legend in the lower-right.

Fig. 6. USGS / NWS River Forecast Center river gauge plot for the morning of 5 October 2015. River gauges experiencing flooding are indicated by the legend in the lower-right.

Fig. 5. Experimental SPoRT-LIS total column relative soil moisture percentile product, valid at 1200 UTC on (left panel) 27 September, and (right panel) 4 October 2015.

Fig. 7. Experimental SPoRT-LIS total column relative soil moisture percentile product, valid at 1200 UTC on (left panel) 27 September, and (right panel) 4 October 2015.  Masked white areas represent water or urban pixels.

An Evaluation of the SFR Product during a mid-December Winter Storm for Northern New Mexico

The Albuquerque NWS recently began receiving an updated NESDIS snowfall rate (SFR) product from NASA SPoRT. We were anxious to see how the updated product performed during our most recent winter storm. A fast moving upper level trough and associated Pacific Front blasted into western New Mexico on the afternoon of Saturday, December 13. The upper low deepened and closed off over New Mexico with wrap around snow impacting northeast New Mexico through mid-day Sunday, December 14.  Ahead of the system, temperatures were very warm with Albuquerque reporting a high of 61 and Santa Fe reporting a high of 57 on Saturday.  The RGB snow-cloud product from 2045Z on Sunday depicts snow cover following the event. Four areas in the state were impacted – the western high terrain, the San Juan and Sangre de Cristo Mountains (mainly west slopes) in north central New Mexico, and extreme northeastern corner of New Mexico. Four yellow ovals mark areas to be discussed in this blog entry. Strong westerly, downslope flow on the backside of this storm system resulted in the snow-free region along the eastern slopes between Taos and Raton.
SnowCloud121414_2045Z

In the loop below, the 0.5 reflectivity mosaic and surface observations show the surface front moving into western New Mexico (left most oval in the snow-cloud product) during the period from 1942Z to 2318Z. In the first image, the winds have shifted to the northwest in Farmington (FMN) and rain is reported as temperatures are too warm to support snow. Note that throughout the loop the Farmington area, especially west and north of the site, there are no radar returns. The Four Corners area has poor to no radar coverage and it is an area where we hope the SFR product will help us. Snow was reported at Gallup (GUP) by 2030Z.

0.5 Reflectivity 1942Z to 2318Z

The SFR product was limited during this initial period, with only one swath covering New Mexico at 2034Z (shown below). This image (obtained from the SPoRT product page) shows that snow is detected in northeast Utah and northwest Colorado, but not in northwest New Mexico.  The Gallup area ended up with about one inch of snow while higher terrain south of Gallup reported two to three inches. While only rain was reported at the Farmington ASOS, the snow-cloud product shows some snow just to the east of Farmington where reports of one-half to an inch of snow was reported.

SPoRT_SFR_121314_2034Z

The next SFR product with coverage over New Mexico had a timestamp of 0338Z (14 December 2014), and is compared to the composite reflectivity image of 0336Z in the image below. Reflectivity is strongest just west and northwest of the Albuquerque ASOS (ABQ), which is reporting rain. The cold front however was moving quickly from west to east toward the ABQ metro area. The strong reflectivity returns to the northwest of Albuqurque are actually bright banding as rain began changing over to snow. The dual polarization hydrometeor classification algorithm showed the rain/snow line shifting quickly eastward. Fifteen minutes prior to this image, rain transition to snow was reported in Rio Rancho, just northwest of Albuquerque. The higher terrain just east of Albuquerque, the Sandia and Manzano Mountains, did receive snow accumulations of two to four inches and the SFR product highlights that area with light rates (blue) of about .02 inches/hour. The Santa Fe area (SAF) is not reporting snow at this time, but is highlighted with the max values of SFR, though snow reports in the Santa Fe area were generally less than 2 inches.  Recall that afternoon temperature were quite warm, making it difficult for snow to accumulate. The SFR product also depicts rates up to .05 inches/hour over the Sangre de Cristo mountains north and east of Santa Fe, where accumulations of 4 to 8 inches were reported. Interestingly, the SFR product is estimating precipitation around Santa Fe when the radar reflectivity pattern and observation do not indicate rain or snow. A portion of this area to the immediate northeast and east of ABQ is beam-blocked by the Sandia Mountains (yellow oval southwest of SAF).

mosaic_Comp_Ref_20141214_0336_SFR_0338Z

A similar comparison is shown for 13 hours later, or around 1655Z on December 15. (Another image was available around 08Z, but is not discussed in the post.)  Note that the SFR product depicts accumulating snow, albeit light, from eastern Taos through all but extreme southern Colfax County. Two stations (KAXX and KRTN) are reporting snow, but radar composite reflectivities do not extend over either location. Snow did accumulate at KAXX, but not at Raton (KRTN) where temperatures hovered right above freezing.

mosaic_Comp_Ref_20141214_1642_SFR_1645Z

Snow that is evident in extreme northeast New Mexico occurred after mainly 16Z and was associated with persistent wrap around precipitation (a SFR product was not available). The SFR product was not used in near real time for this event but was re-examined only a short time thereafter. However, the product did validate that we will indeed be able to complement radar void coverage areas in an operational forecast environment using polar-orbiting satellite imagery. This example will also serve to highlight potential product applications, advantages, and disadvantages for forecaster training prior to the upcoming NESDIS evaluation period.

Example of Limb Effects on NESDIS SFR Product

SPoRT continues to work with select NWS WFOs in evaluating the NESDIS SFR product.

One thing to take into consideration when using data from “whisk-broom” instruments on polar-orbiting satellites, such as the Advanced Microwave Sounding Unit (AMSU) used to generate the SFR product, is that data at the edge of the swath (i.e. limb) may provide misleading or erroneous observations. As the instrument scans farther from nadir, it is looking through more of the atmosphere, creating both a bigger observation field of view (i.e., larger pixel) and having the signal attenuated by more atmospheric constituents (e.g., in-cloud and falling snow).  As a result, when interpreting the SFR product, it is important to look for the extent of the swath (outlined in gray in the product in AWIPS) to determine whether the observed SFR is going to be limited by these limb effects.

Let’s take a look at an example over the NY Tri-State area for the post Super Bowl snow event.  In the first image, from Metop-A valid at 1458 UTC, there is a large area of snowfall across the area.  The heaviest SFRs appear to be around 1.2-1.5 in/hr (when multiplying the liquid equivalent by 10) across central and southern New Jersey.  However, an hour later (see second image from Metop-B valid at 1554 UTC), the shape of the heaviest SFR has expanded north and west and there are now readings over 2.0 +in/hr.  Other areas that had a SFR of less than 0.5 in/hr in the 1458 UTC image, appear to have a SFR of around 1.0 in/hr just an hour later, which is a large jump in intensity.

While it is certainly possible that the snow evolved and intensified in less than an hour, it is more likely that instrument limb effects are likely to blame for the larger SFRs in the second image.  Make sure to check that swath edge when using polar-orbiting satellite data!

NESDIS SFR Product from 1458 UTC on 3 February 2014 showing snow detected near nadir for Metop-A

NESDIS SFR Product from 1458 UTC on 3 February 2014 showing snow detected near nadir for Metop-A.

NESDIS SFR Product from 1554 UTC on 3 February 2014 depicting what are likely erroneous higher intensity SFR values along the swath edge from Metop-B.

NESDIS SFR Product from 1554 UTC on 3 February 2014 depicting what are likely erroneous higher intensity SFR values along the swath edge from Metop-B.

NESDIS SFR Captures Central and NE Alabama Snow Event

SPoRT continues to work with select NWS WFOs in evaluating the NESDIS SFR product.

A rare winter storm impacted much of the deep South Tuesday morning and afternoon.  Areas of Central and Northeastern Alabama only received a couple of inches of snow, but this was enough to cause major headaches as roadways iced over resulting in highways across Alabama and Georgia being shut down, stranding thousands of motorists.  Most reports from late Tuesday morning indicated that the worse of the winter weather was falling south of Cullman, AL, through Birmingham, AL to Montgomery, AL and then eastward into areas like Fort Payne, AL.  I-65 north of Birmingham and I-20 east of Birmingham were particularly troublesome in the state of Alabama.

The AWIPS image below depicts the SFR product in AWIPS with overlaid interstate highways.  This image was taken from the AMSU on Metop-B at 1618 UTC (10:18 local Alabama time) right about the time when the heaviest snow was impacting the state.  The SFR Product indicates liquid water equivalent precipitation rates between 0.04 and 0.08 in/hr, which if you multiply by a factor of 10 to get the solid snowfall rate equates to around 0.4 and 0.8 in/hr.  Snowfall totals across this region were generally in the 1-3 inch range, so the rates that were detected by the product were consistent with what actually fell.

NESDIS SFR Product from 1618 UTC (around 10:00 A.M. Central) on 28 January 2014

NESDIS SFR Product from 1618 UTC (around 10:00 A.M. Central) on 28 January 2014

SFR product performance during snow across WFO RLX’s forecast area.

All,

I have attached a screen capture of the SFR product from 1024 UTC on 1/21/14.  The label on the image is wrong.  It states the units of the product are in/hr.  But they are actually mm/hr.

During this time, we were having widespread light to moderate snow as an upper level disturbance moved across our forecast area.  Reports around 2 inches of snow were common around the time of the product.   We had received reports of snow coming down around an inch per hour.  The maximum SFR detected in the product was 1.6 mm/hr…or 0.06 in/hr.  Using a ratio of 15:1 yields a maximum snowfall rate around 0.9 inches per hour.

While we had several surface observations from which we could estimate precipitation rates, our WSR-88D was not operating correctly.  The legacy precipitation were okay.  But the Dual Pol precipitation products were not totally reliable due to equipment issues.  So the additional information from the SFR product should have helped estimate the precipitation rates.

SFR_1024Z_012114