14 April 2012 - Unique applications of NDVI and GOES dryness products

Bob Rabin from NSSL/UW-CIMSS provided me with a couple interesting images late last week in preparation for the upcoming events on Friday/Saturday. I thought I would share them here with you all to get a new perspective on a product that we used extensively within the Fire Weather Experiment last Fall to monitor vegetation and surface moisture. Bob Rabin gathers and generates NDVI composite imagery as well as GOES derived surface dryness values that we provide within SPC operations in support of their fire weather forecast desk. The products were mainly intended to monitor for dry vegetation and anomalously dry surface conditions that would be a potential hazard for fires. In these examples, Bob pointed out that the NDVI and GOES surface dryness products were both picking up on a very well defined dry / moist boundary across much of KS, OK and TX (see images below).

Continental US NDVI composite from 9 April 2012. Green areas indicate regions with significant green vegetation cover, while yellow and brown regions indicate decreased green vegetation cover. Note the pronounced gradient in green vegetation along a line extending from western KS and OK into central TX.

Continental US GOES surface dryness composite from 11 April 2012. Green areas indicate regions with significant surface moisture mesurements, while yellow and red regions indicate dry surface measurements. Note the pronounced gradient in surface moisture along a line extending from western OK into central TX. It would be interesting to examine how these surface moisture boundaries interacted to potentially enhance convective initiation set up by circulations due to differential heating. In fact, one SPC forecaster noted this in a Mesoscale Discussion issued on the 14th prior to convective initiation along the TX/OK border. In addition, it would be interesting to examine the interaction these surface moisture boundaries have on the evolution of the dryline. It is exciting to see applications of satellite imagery used to detect these features not easily observable from other systems in such a constant and relatively highly spatial manner.

14 April 2012 - Simulated Band Difference

Part of the NSSL-WRF simulated satellite imagery that we get from CIRA includes a band difference unique to GOES-R that we began looking at during last year's Spring Experiment. One of the advantages of simulating satellite data from a model is that we have the opportunity to produce channels that we don't have currently, and we take full advantage of this by producing all 9 of the non-solar GOES-R IR bands. The 10.35 micron channel is a very clean window, and thus is very sensitive to surface temperature. The 12.3 micron channel however is sensitive to low- and mid-level water vapor. As moisture moves into a clear pixel area, the 12.3 micron brightness temperature will decrease, whereas the 10.35 micron temperature should stay the same. When this occurs, the 10.35-12.3 micron channel difference will become strongly positive and indicates areas of moisture convergence or pooling, which can lead to destabilization and subsequent convective initiation.
Unfortunately, because the imagery is generated by a numerical model, it is a) not an observation and b) only available on an hourly timescale. However, we can use the imagery generated from the model as a experimental tool to demonstrate some of the unique things we can do once we have the increased spectral resolution of the GOES-R Advanced Baseline Imager (ABI). In this example from the 14 April 2012 outbreak the 10.35-12.3 micron channel difference is useful in identifying the evolution of the dryline across western KS, OK and TX from 1900 UTC on the 14th to 0100 on the 15th (see images below). As moisture converges at the surface, the difference becomes more positive. These positive values show up as yellow, orange and red on the images below. The edge of the dryline is easy to detect and follow using this simple band difference within the NSSL-WRF. It will be very interesting to see observations of this band difference, and other imagery techniques such as RGBs, every 5 minutes over the continental US once we have the GOES-R ABI available to us.

14 April 2012 - Sounder Airmass RGB

Well, it's not quite the Spring Experiment yet, but we did have a significant severe weather event over the plains this past weekend that I figured would be a good opportunity to capture some of the GOES-R Proving Ground products that we receive at the Storm Prediction Center and Hazardous Weather Testbed. We will start with the GOES Sounder Airmass RGB that is provided to us by CIRA and NASA SPoRT. RGB simply stands for Red-Green-Blue, which is a composite image created by combining three separate channels or channel differences into one image. This technique helps us identify specific features in the atmosphere without the use of complex derived products. The airmass RGB is a combination of thermal infrared, water vapor and ozone channels that help us identify regions of warm and moist versus cold and dry airmasses, spin in the atmosphere and jet streaks. The airmass RGB has been used extensively over Europe using the Meteosat Second Generation satellite, which has similar spectral channels to what will be available on the GOES-R Advanced Baseline Imager (ABI). Current GOES imagers do not contain the spectral bands necessary to generate this product, but we are able to simulate the RGB using the GOES sounder, which does have some similar channels to the ABI. Unfortunately this data only arrives once hourly from the sounder, but when GOES-R is launched, we will be able to create this product every 5 minutes over the continental US.
Below is a time-series of the SPC's outlooks from day 7 up until the event on Saturday 14 April 2012 from the SPC Facebook page. You can see clearly that the SPC had a good handle on the threat a week in advance and that the threat area depicted on day 1 well outlined the events that occurred in NE, KS, OK and IA. Preliminary storm reports from the SPC website indicate 135 tornado reports (likely will end up being about 75 individual tornadoes following official surveys) occurred during this event, with what appears to be several long tracks across OK and KS.

If we take a look at the sounder airmass RGB product at 12 UTC on 14 April 2012 (top image below), we can see a strong center of circulation over CA/AZ/UT/NV with indications of a significantly lowered tropopause, associated jet streaks and high PV (red hues) within the circulation and extending along the Pacific coast up into Canada. This can be confirmed by overlaying the tropopause pressure from the RUC analysis (middle image below) or the the 500 mb heights and vorticity (bottom image below).

If we move forward to 15 UTC (image below), just prior to initiation in KS, we can begin to see evidence of a moisture boundary (blue-to-green hue gradient) setting up along a line extending from NW KS near the KS/NE border down into the OK panhandle and down into west TX. Initiation in southern KS and down through the OK panhandle into TX would occur along a dryline located in this area, which is likely what the airmass RGB is picking up at this time.

Moving forward to 02 (top image below), 03 (middle image below) and 04 UTC (bottom image below) on 15 April 2012, the moisture gradient feature really begins to tighten up and better define itself as the dryline continued to evolve and the Pacific cold front approached, initiating a line of storms extending into central TX.