A couple more RGB observations this week…

Ok, while I have a mintue I thought I’d make another post on observations with RGB imagery in the past few days.  The first image below is essentially a followup to Brian Guyer’s excellent post on the 24th.

Figure 1. MODIS RGB Nighttime Microphysics image valid May 24, 2012 at 0835Z. The burn scar appears as the red “heart-shaped” area in SW New Mexico.

I wondered what the Nighttime Microphysics RGB would show for this burn scar, given the heavy use of the 10.8 µm channel in this product’s “recipe”.   Notice the burn scar shows quite well, and is the pink, nearly heart-shaped area to the NW of KSVC.  This image was taken by the MODIS sensor at approximately 0835Z, or 0235 local (MDT).  I’ll have to admit I was a little perplexed, as I expected this area to be warmer than the surrounding terrain.  However, the decrease in blue contribution in this area (the blue in the RGB ”recipe” corresponds to the 10.8 um channel) would indicate a relatively colder surface.  Perhaps the radiational cooling was more efficient here?  Perhaps one of you folks with either more experience with RGBs or burn scars could help here.

Now, for a change as I take a look at another fog example…this time from my own forecast area (HUN).  The first image below is the MODIS RGB Nighttime Microphysics image valid May 23rd at 0340Z (1040 local CDT).

Figure 2. MODIS RGB Nighttime Microphysics image valid May 23, 2012 at 0340Z.

At first it may be a little difficult to see, but notice the aqua colors that show up in the extreme NE part of AL, but mainly in portions of southern Tennessee that are highlighted by the yellow circles.  This color in the Nighttime RGB imagery is typically indicative of the presence of fog and/or low clouds.  The southern circle represents the narrow valley locations along and near the Tennessee River, while the nothern circle represents fog and low clouds in the narrow Sequatchie Valley.  This valley area often acts as a region of local cool air production and the development of early fog.  As the cool air flows southward and gradually downhill along the TN River channel, fog development will spread into Jackson County.  The RGB image below, shows the relatively early development of fog in these critical areas.  However, notice the presence of slightly deeper convective clouds in the area, as indicated by the more yellow and tan colors.  An upper low was still exiting the area and lapse rates above the developing nocturnal inversion were enough to generate the convective clouds.  In the standard 11-3.9 um imagery below, it is nearly impossible to distinguish between the convective (but still fairly shallow) clouds and the low clouds/fog.

Figure 3. GOES-MODIS hybrid valid May 23, 2012 at 0345Z.

The RGB imagery here shows great utility in detecting fog under these difficult circumstances.  The next RGB image below shows the eventual development of fog that morning by 0800Z.  Here, the fog can be seen to have clearly spread throughout the valley areas of Jackson County (northeasternmost county in AL).  The Scottsoboro observation indicated visibilities below one quarter mile at this point.

Figure 4. MODIS RGB Nighttime Microphysics image valid May 23, 2012 at 0752Z. The aqua colors in NE Alabama, southern TN, and NW Georgia indicate fog/low clouds that have spread rapidly through valley areas overnight.

Here again, the RGB imagery show superior utility in the early detection of fog over traditional imagery that can have benefit to operations.

Whitewater-Baldy Fire Complex Hotspot Comparison

The Whitewater and Baldy Fires within the Gila Wilderness of New Mexico exhibited extreme fire behavior during the afternoon of Wednesday May 23, 2012.  The Baldy Fire started on May 9th and the Whitewater Fire on May 16th both due to lightning strikes.  These two fires merged into one large fire complex on the 23rd while growing nearly 50,000 acres in just one afternoon.  The 1km MODIS 3.9micron hotspot imagery comparison between Tuesday, May 22nd and Wednesday, May 23rd showed this significant growth.  Figure 1 is a graphicast issued on the evening of the 23rd using the comparison to show the incredible growth.  Forecasters were able to approximate the size of the fire by comparing the imagery to burn scars from 2011.  The Las Conchas Fire burned nearly 150,000 acres to the west of Los Alamos, becoming the largest wildfire in New Mexico history.  This fire was easily close to half that of the burn scar displayed in the graphicast in Figure 2.  The following morning inciweb.org indicated the fire had grown to near 50,000 acres.  By Thursday afternoon the 24th the fire continued to grow to nearly 70,000 acres.  The Miller Fire burn scar east of the current fire in Figure 2 burned just over 88,000 acres.  This looks very close to the latest 1km MODIS hotspot image.  Given the current long duration critical fire weather event expected through Saturday it will be unprecedented to see the Whitewater-Baldy complex grow to beat the state record set just last year.

Figure 1. Graphicast of MODIS Comparison.

Figure 2. 2011 burn scar comparion to the current Whitewater-Baldy complex as seen from MODIS.

Follow-up to HWT CI Event in TX Panhandle: Radar Loop

The figure below shows a radar loop of the supercell for 1.5 hours leading up to the time of the tornado.  Notice how the supercell to the WNW of Amarillo, TX develops nearly in place, along the NW edge of the moist soil patch as documented in the previous blog post on this topic.  Prevailing low-level flow was from the SE, which may have led to maximum convergence to the NW of the moist patch, given a “sea breeze-like” circulation due to differential surface heating.  I plan to make some WRF model simulations using the NASA Land Information System to further explore the role of soil moisture patterns in developing the convection in west Texas on this day. (Click twice on the image below to see the radar animation.)

FIG 1. One and a half hour animation of base reflectivity over the Texas Panhandle, valid from 0000 UTC to 0130 UTC 22 May 2012.  Credit: Plymouth State University Weather Center web page.

Passive Microwave Composites of Tropical Storm Bud

Following up from our previous post regarding Aletta, today’s pass of DMSP F-18 (1535 UTC) was nicely centered over Tropical Storm Bud, located southwest of Mexico in the eastern Pacific.  The 1500 UTC advisory suggests maximum sustained winds of 65 mph and northwest motion, with a northern turn expected.  It may progress to develop into a hurricane later today or tonight.  Below are current snapshots of RGB passive microwave composites provided to SPoRT via collaboration with the Naval Research Laboratory, converted to use in GEMPAK and N-AWIPS.  The 89 GHz image (top) includes estimates of the observation time (1530-1535 UTC) and the satellite instrument (DMSP F-18).  These composites demonstrate how passive microwave data can help to identify structures of the tropical cyclone that might otherwise be obscured by thick cirrus overhead.  SPoRT plans to collaborate with NCEP groups to test and evaluate these products in N-AWIPS in support of our NRL collaboration, as these products have been very popular when provided by the NRL Tropical Cyclones web page.

91 GHz RGB composite image from DMSP F-18 SSMI/S sensor over Tropical Storm Bud at approximately 1530 UTC on May 23, 2012.

As in the image above but for the 37 GHz channels.

Convective Initiation Event during the HWT Experimental Forecast Program

It’s Jonathan Case reporting from the Hazardous Weather Testbed (HWT) Experimental Forecast Program (EFP) in Norman, OK.  At the EFP this year, teams of researchers and operational forecasters are coming together to examine cutting-edge, convection-allowing ensemble numerical model output, unique diagnostic model fields, and experimental observation platforms to forecast convective initiation and severe weather events.  Each day, the participants are divided into sub-groups with one group forecasting severe weather over a pre-selected focus region while the other group forecasts convective initiation and coverage in the same general sub-domain.

On Monday, the focus region was over eastern New Mexico and the Texas Panhandle.  Model guidance was pretty good at indicating storms developing over the higher terrain of NM, and then conglomerating into a SE-moving mesoscale complex over the TX Panhandle after 00z 22 May. During verification activities Tuesday morning, an interesting feature was seen in the visible satellite imagery over the western Texas Panhandle in the afternoon and early evening hours.  A small area had a dearth in cumulus cloud development a little east of the NM border, as highlighted in FIG 1.  At the time of the visible image in FIG 1 (2245 UTC), convective storms began developing on the NW edge of this cloud-free “patch”, ahead of the developing convection in NM.  This convection intensified with time, as seen in the radar image in FIG 2.  In fact, a tornado was reported at 0137 UTC 22 May not far from this location (http://www.spc.noaa.gov/climo/reports/120521_rpts.html), as the convection on the NW side of the “patch” continued to intensify into an isolated supercell over the next 2 hours (not shown).

FIG 3 is a 500-meter true-color image from MODIS, centered over the TX Panhandle at 1729 UTC 21 May, before cumulus cloud development occurred on the periphery of the “patch”.  One can distinctly see the darker brown color of the “patch” compared to the lighter brown color of the surrounding landscape.  Further investigation reveals that the darker brown color of the “patch” correlated very well with high antecedent rainfall from the previous day (FIG 4; 24-hour rainfall of ~0.5″ to 2.0″, ending the morning before the event of interest).

What likely happened is that the previous day’s rainfall moistened the top soil layer over this patch, thereby reducing the overall surface albedo making the land surface appear darker in the MODIS true color image.  Because more incoming shortwave energy was partitioned into evaporating soil moisture, the surface heated more slowly than the surrounding landscape.  This differential heating appears to have led to a small-scale “sea-breeze-like” circulation, enhancing the cumulus cloud development on the periphery of the moist patch by late afternoon.  Additional research is required to determine whether this moist patch ultimately contributed to the tornado occurrence.

FIG 1. GOES visible satellite image at 2245 UTC 21 May 2012. Red label indicates location of patch where a lack of cumulus cloud development occurred.

FIG 2. Base reflectivity image over the Texas Panhandle at 2259 UTC 21 May 2012. Red outline indicates the convective cell developing on the NW edge of the cloud-free patch.

FIG 3. MODIS true color image at 500 m resolution, centered on the Texas Panhandle, valid at 1729 UTC 21 May 2012.

FIG 4. Stage IV precipitation analysis over the Amarillo, TX WFO county warning area for the 24-hour period ending 1200 UTC 21 May 2012.

NRL Passive Microwave Products Available for N-AWIPS: Tropical Depression Aletta

Tropical Depression Aletta is slowly churning through the eastern Pacific, currently located southwest of Baja California in the Eastern Pacific.  As of 0800 PDT, the latest forecast advisory estimates peak winds of 35 mph and the storm is expected to weaken.  However, the presence of the storm is a great opportunity to feature the results of a collaboration between the Naval Research Laboratory and the SPoRT Center that has facilitated the integration of NRL passive microwave products into the N-AWIPS visualization system (which is based upon GEMPAK capabilities).  Below are some comparisons between graphics provided by the NRL Tropical Cyclones page.  SPoRT has developed display capabilities for NRL single channel and multispectral (RGB) passive microwave products along with an estimate of the observing time based upon the satellite orbit.  These images were obtained from the DMSP F-17, although data from the full constellation of DMSP satellites and NASA’s Tropical Rainfall Measuring Mission (TRMM) are also available.  Since these satellites have slightly different orbits, each snapshot of the storm is provided to SPoRT by NRL, and data are processed for display in N-AWIPS and distributed to partnering NCEP locations.

Low frequency (37 GHz) passive microwave RGB composite of Aletta at approximately 1305 UTC, based upon observations form DMSP F-17.

Equivalent to image above, produced by SPoRT from NRL data and configured for display in N-AWIPS (GEMPAK) for dissemination to partnering NCEP Centers.

High frequency (87-91 GHz) image of Aletta provided by the NRL Tropical Cyclones page.

Equivalent image processed by SPoRT for dissemination in N-AWIPS.

 

Although the zoom factor is slightly different between the NRL image and the N-AWIPS examples, the display of the data in N-AWIPS will allow forecasters to have greater control in manipulating the imagery and collocating satellite observations with other analyses available within their local systems.

RGB Products Show Enhanced Utility for Fog Monitoring

It was a “typical” post-rain foggy morning for portions of the Appalachian region this morning.  When looking at the suite of SPoRT satellite products this morning, I noticed the usual fog in the GOES and MODIS imagery, but what caught my attention was the early detection of fog by the Nighttime Microphysics RGB product.  Image 1 below shows the GOES-MODIS 10.8-3.9 µm product currently created by SPoRT and available through the Southern Region LDM to collaborating offices.

Image 1. GOES-MODIS 11-3.9 µm hybrid image valid May 16, 2012 0331Z.

In this standard color scheme, the yellow colors generally correspond to low clouds and/or fog while the blue colors represent high clouds.  Notice that low clouds and fog can be seen in portions of the Smoky Mtns region in North Carolina and East Tennessee, from around Waynesville, NC to Johnson City, TN.  Also, areas of fog and low clouds can be seen in portions of West Virgina and western Virgina.  Now, take a look at the Nighttime Microphysics RGB below (Figure 2)…

Figure 2. SPoRT Nighttime Microphysics RGB product valid May 16, 2012 0334Z.

Notice in the RGB image that the areas of fog mentioned above show up as the light aqua colors. However, notice the very light aqua colors that show up in the Cumberlands and Alleghenys to the west. In this area, generally from Livingston, TN northeastward through eastern Kentucky and into western and central portions of West Virgina, fog shows up fairly well in the narrow valleys of the Cumberland/Allegheny Plateau region. I was surprised to see this much difference between the two products, but the enhancement of the RGB product and the ability to discern the shallow fog and low clouds from the terrain background can be helpful to forecasters. This is possible with the more complex “recipe” of the RGB Nighttime Microphysics product (Red: 12.0-10.8 µm, Green: 10.8-3.9 µm, Blue: 10.8 µm). In this case, it would have alerted the forecasters to the early development of valley fog. Given a continuation of favorable conditions (which did occur), the fog would likely spread and become more dense with time overnight. See the RGB image below valid at 0746Z, which indicates the rapid expansion of the fog.

Notice the aqua colors have spread markedly across the valleys of this region over the ensuing 4 hour period. Most of the NWS offices with forecast responsibility in this region (Morristown, Blacksburg, Jackson (KY)) issued dense fog advisories during the early morning hours for their affected counties. These advisories came several hours after the early detection of fog by the RGB Nighttime Microphysics product, and most forecasters are not going to issue these products until its evident dense fog is going to form anyway (and speaking from experience, that doesn’t often come easily). Of course, most of these offices do not get these types of satellite products yet, but they do certainly show utility in the future in these types of cases. With pattern recognition, the early detection of fog amidst favorable conditions might suggest that dense fog would eventually form. Had forecasters been able to see the locations of fog development relatively early (as shown in the 0334Z RGB product), it could have increased their confidence in the development of widespread dense fog, which could have been reflected properly in Area Forecast Discussions or other advisories or statements during the late evening.