Archive for July, 2007

Rapid warming of southern Lake Michigan water temperatures

Monday, July 23rd, 2007

AWIPS MODIS SST image

An AWIPS image of the MODIS Sea Surface Temperature (SST) product (above) revealed a pocket of warmer SST values (72-75º F or 22-24º C, orange enhancement) over the mid-lake waters of southern Lake Michigan during the afternoon (around 19:41 UTC, or 2:41 PM local time) on 23 July 2007.

An animation of daily MODIS SST images (below) indicates that this particular warm water feature was not evident on afternoon MODIS SST imagery over southern Lake Michigan on 20 July or 21 July, but the water temperatures did begin to increase over that general region on 22 July.

MODIS SST images (Animated GIF)

A time series plot (below) of the air temperature and water temperature from the southern Lake Michigan Buoy 45007 (data courtesy of the National Data Buoy Center) shows that water temperatures at that location (along the northern fringe of the warm SST feature) warmed at a rate of nearly 1º F per hour during the afternoon hours on 23 July (with a total increase of 9º F or 5º C in the 12-hour period from 12:00-00:00 UTC). It is interesting to note that the Buoy 45007 water temperature (blue; measured at 0.6 meter below the site elevation) warmed more quickly than the Buoy 45007 air temperature (red; measured at 4 meters above the site elevation) during the daytime hours (12 GMT to 00 GMT) on both on 22 July and 23 July. Relatively light winds (4 knots or less) and low wave heights during the morning and afternoon allowed for such a rapid warming of the lake’s “skin temperature” (a similar diurnal change in SST values over the Gulf of Mexico and the western Atlantic Ocean — as much as 3º K (5º F) in regions of light winds — was previously reported using GOES-8/GOES-9 IR satellite data [Wu, Menzel, and Wade, 1999]). The maximum MODIS SST values of 77º F (25º C) seen on 23 July were the warmest observed over southern Lake Michigan during the 8-day period from 17-24 July.

Buoy 45007 air and water temperature

GOES-13 vs. GOES-12 visible channel

Friday, July 20th, 2007

GOES-13 visible image

The GOES-13 satellite (launched in May 2006) has been brought out of on-orbit storage for a few weeks of testing and evaluation. A comparison of 1-kilometer resolution visible channel imagery (centered on Sioux Falls, South Dakota) from GOES-13 (above) and GOES-12 (below) demonstrates the fact that over time there is some in-orbit degradation of the GOES visible detectors. Note that the network of cities, towns and highways can be seen in the GOES-13 visible image above, especially across northwestern Iowa (in particular, Highway 60 which runs southwest to northeast: Google maps) and southwestern Minnesota — these towns and roads show up due to the contrast between the higher albedo concrete of the towns and road surfaces (and the adjacent ditches/medians) and the lower albedo of the surrounding fields of dense, mature corn crops. These features were less apparent in the GOES-12 visible image below (GOES-12 was launched in July 2001, and has been the operational GOES-East satellite since March 2003).

GOES-12 visible image

Part of the difference in the appearance of the scene can be explained by examining a plot of the spectral response function of the visible channels on GOES-12 and GOES-13 (below; thanks, MatG!): the sharper cutoff for wavelengths beyond 0.7µm on the GOES-13 visible channel (red line) makes it less sensitive to the signal from the mature corn crops (green line), allowing greater contrast between the thick vegetation of the agricultural fields and the more sparsely vegetated cities, towns, and highway corridors. Also of significance is the fact that Highway 60 in Iowa was in the midst of a major reconstruction project to expand it from a 2-lane highway to a 4-lane divided highway, making for a wider non-vegetated space that more easily shows up on the 1-km resolution GOES-13 visible imagery.

GOES-12. GOES-13 spectral response functions

A 250-meter resolution “true color” RGB composite image from the polar-orbiting Terra MODIS instrument (below) shows even better detail, with the roadways, cities, and towns standing out very well against the surrounding background of dark green corn crops. Even the square grid network of minor county roads (spaced at 1 mile intervals) can be seen in the MODIS image.

Terra MODIS true color image

Wildfires in Idaho/Nevada

Friday, July 20th, 2007

AWIPS GOES + MODIS IR images

Increasing wildfire activity was noted across parts of the Great Basin region of the western US on 20 July 2007 — according to the National Fire Interagency Center, there were 43 large fires burning in the states of Idaho and Nevada on that day (NOAA HMS fire and smoke product). AWIPS images of the GOES and MODIS 3.9µm/3.7µm IR channels (above; top 2 image panels) and the 10.7µm/11.0µm (”IR window”) channels (above; bottom 2 image panels) showed a cluster of active fires straddling the Idaho/Nevada border around 05:30 UTC. The 4-km resolution GOES 3.9µm IR image (upper left panel) showed 3 separate areas of somewhat warm “fire hot spots” (dark grey to black pixels), while the 1-km resolution MODIS 3.7µm IR image (upper right panel) revealed a larger number of more intense “fire hot spots” (black to yellow to red pixels).

Since the areal coverage and intensity of the GOES hot spots was diminishing during the nighttime hours, the GOES-derived Wildfire ABBA product (below) did not indicate any fire activity along the Idaho/Nevada border region around 05:30 UTC; therefore, the 1-km resolution MODIS 3.7µm IR imagery shown above would have been the only satellite-based diagnostic of fire activity at that particular time.

GOES-11 Wildfire ABBA product