GOES-13 Sounder 10.7 µm imagery without and with Processing Software changes (click to enlarge)

The GOES-13 (GOES-East) Sounder instrument has been experiencing data anomalies that manifest themselves as missing pixels (link, link). These errors occurred because of slight fluctuations in the speed of the sounder filter wheel, resulting in a time offset. Processing software expects the data to be present at a certain time, but because of the filter wheel speed fluctuations, data were not present when expected. All 19 spectral bands on the GOES-13 Sounder were affected.

Missing data can now be reclaimed using a modified version of the SPS (Sensor Processing System), the ground software that makes the GVAR data stream. GOES Engineers have been testing this software change. The test software modifications properly handle the slight differences in the timing of the data. As of this time, a date has not been slated for operational implementation by NOAA NESDIS.

An example with all 19 bands of the GOES-13 Sounder is shown below, with the current pixel drop-outs (top) and after the software changes were applied (bottom) as part of off-line testing (image toggle).

GOES-13 Sounder imagery (all 19 bands) [click to enlarge]

GOES-13 Sounder imagery (all 19 bands) produced with new processing software [click to enlarge]

View only this post Read Less

Landsat-8 0.59 µm panochromatic visible image of the Washington State landslide site

Kudos to Russ Dengel of the SSEC RealEarth web map server development team for spotting this: a relatively cloud-free overpass of the Landsat-8 satellite which revealed the site of the massive landslide/mudslide near the small town of Oso in northwestern Washington State (north of Seattle). The animation shown above was made using RealEarth to zoom in with the Google Maps base layer, and then toggle between the base map and an overlay of 15-meter resolution Landsat-8 0.59 µm (Band 8) panochromatic visible imagery at 19:03 UTC or 12:03 PM local time on 23 March 2014. It can be seen that debris from the landslide — which occurred a day earlier — covered one mile of State Road 530, cutting off access to the town of Darlington (located east of the landslide site); it also blocked the North Fork of the Stillaguamish River, leading to fears of localized flooding both upstream and downstream of the landslide site.

The Landsat-8 visible image is shown below. The landslide was blamed on ground saturation due to heavy rainfall in the region over the past month (30-day total rainfall | depature from normal).

Landsat-8 0.59 µm panochromatic visible image

===== 01 April Update =====

Landsat-8 0.59 µm panochromatic visible image on 01 April

14 days later, there was another overpass of the Landsat-8 satellite; it could be seen on the 0.59 µm panochromatic visible image (above) that the mudslide still covered a significant portion of State Road 530. The 1.61 µm near-IR image (below) revealed that the mudslide had altered the course of the North Fork of the Stillaguamish River, and highlighted areas where some localized flooding was occurring due to a widening of the river (water is a strong absorber at the 1.61 µm wavelength, so it appears dark on the near-IR image).

Landsat-8 1.61 µm near-IR image on 01 April

View only this post Read Less

Suomi NPP Sea Surface Temperature product, 0727 UTC 21 March 2014

Suomi NPP VIIRS data are being used to compute sea-surface temperature (SST), and those fields are now being input into AWIPS for evaluation. The Suomi NPP images as displayed in AWIPS are labeled as MODIS SSTs (this was done to speed the injection process into AWIPS; however, note that the labels on VIIRS images for this blog post were modified to display the correct satellite source). Even though the labels are the same (both MODIS and VIIRS products are labeled “MODIS Sea Sfc Temperature” in AWIPS), the data sources are different, and the user can learn to identify the data being used.

A user can match the time of the image to overpass times for the Aqua, Terra or Suomi NPP satellites. Overlaying a different Suomi NPP image, for example 11.45 µm IR brightness temperature, below, that covers the same geographic region will also tell the user which data source — from VIIRS or MODIS — is being used to construct the SST product.

VIIRS-based SSTs and 11.45 µm IR brightness temperatures (click to enlarge)

A MODIS-based SST from approximately the same time is shown below. The VIIRS swath is much wider than the MODIS swath. This will always be true.

MODIS and VIIRS-based SSTs at ~0730 UTC 21 March 2014 (click to enlarge)

Values for MODIS SSTs and VIIRS SSTs are similar. In general, the cloud-clearing with Suomi NPP VIIRS is more accurate, meaning there are more clear pixels with the Suomi NPP data and therefore more SST pixels. Note in particular differences in the strong temperature gradient along the edges of the Gulf Stream where MODIS algorithms mistakenly flag pixels as cloudy.

[Added: This animation cycles through the SSTs and the Window Channel IR images from both VIIRS and MODIS]

2 days earlier, a nighttime/daytime Suomi NPP VIIRS SST comparison on 19 March — magnified to provide a closer look at the Gulf of Mexico (below) — revealed intricate structure associated with the Loop Curent (the large darker red feature, with SST values around 80º F), as well as other small-scale eddys in the surrounding Gulf waters.

Suomi NPP VIIRS SST images at 08:03 UTC and 19:27 UTC on 19 March

View only this post Read Less

GOES-13 0.63 µm visible channel images (click to play animation)

McIDAS images of GOES-13 0.63 µm visible channel data (above; click image to play animation) showed two very large blowing dust features during the afternoon hours on 18 March 2014: one moving southward out of southeastern Colorado, and another moving eastward out of New Mexico across Texas and Oklahoma. Winds gusted to 75 mph in southeastern Colorado, and 60 mph in the Texas Panhandle; surface visibilities were reduced to near zero at times at some locations. As a result, some highways were closed in southeastern Colorado. Another significant blowing dust event had impacted much of this same region one week earlier.

GOES-13 0.63 µm visible channel images with METAR surface reports (click to play animation)

Taking a closer look at the large southward-moving dust plume with AWIPS images of GOES-13 0.63 µm visible channel data (above; click image to play animation) and GOES-13 3.9 µm shortwave IR channel data (below; click image to play animation), it can be seen that a large wildfire began to burn around 19:15 UTC (2:15 PM local time) near the Texas/Oklahoma border as the strong winds associated with the leading edge of the dust storm moved through. The hot fire (yellow to red to black pixels on the shortwave IR images) produced pyrocumulus clouds and a smoke plume that rose above the top of the dust layer, high enough to cast a shadow due to the low sun angle at the end of the day. Note from the distance scale on the lower left portion of the visible images that this “wall of dust” was at least 300 miles wide.

GOES-13 3.9 µm shortwave IR images with METAR surface reports (click to play animation)

The CLAVR-x POES AVHRR Cloud Top Height product (below) indicated that the top of the southward-moving dust layer was generally 1-2 km above ground level (cyan to light green color enhancement).

POES AVHRR Cloud Top Height product

Terra and Aqua MODIS true-color Red/Green/Blue (RGB) images visualized using the SSEC RealEarth web map server (below) revealed that the composition of the dust from the 2 source regions took on a different appearance, indicating that the nature and composition of the aerosols were different.

Terra and Aqua MODIS true-color RGB images

AWIPS images of the Terra and Aqua MODIS “reverse absorption” 11-12 µm IR brightness temperature difference product (below) offered another method of identifying the areas where the airborne dust was the most dense.

Terra and Aqua MODIS 12.0-11.0 µm IR brightness temperature difference product

Additional satellite images from this event can be found on the Wide World of SPoRT and RAMMB: GOES-R Proving Ground Blog sites.

===== 19 March Update =====

IDEA-I forecast aerosol trajectories (click to play animation)

The IDEA-I MODIS Aerosol Optical Depth product indicated that much of the airborne dust remained over southern and eastern Texas on 19 March. Forecast forward trajectories (above; click image to play animation) suggested that some of this dust would get recirculated back northward across western Texas, and eventually move over Kansas in 24-48 hours.

View only this post Read Less