MTSAT-1R Band 5 Anomaly

September 18th, 2009 |

MTSAT-1R shortwave IR image

Night-time MTSAT-1R 3.75 µm shortwave IR image (14:30 UTC)

A curious image artifact was noted on MTSAT-1R 3.75 µm shortwave IR imagery, in the form of a very cold “false eye” appearing  just to the right of the actual eye of Typhoon Choi-Wan (17W) in the western North Pacific Ocean . The “ghost” of the eye exhibited a satellite radiance of zero (very cold, appearing bright white on the image), and was offset from the true eye by 16 pixels in the horizontal and 1-2 pixels in the vertical. This false eye was most apparent on MTSAT-1R imagery during the local night-time hours, as was seen on 16 September 2009 at 14:30 UTC (above) and 15:30 UTC (below).

MTSAT-1R shortwave IR image (15:30 UTC)

Night-time MTSAT-1R 3.75 µm shortwave IR image (15:30 UTC)

This false eye artifact was also evident during local daytime hours, but the “ghost” did not exhibit zero radiance — the feature could be seen better on the shortwave IR imagery once a contrast stretch  enhancement was applied (below). The offsets of the false eye were the same as seen during local night-time hours.

MTSAT-1R shortwave IR (original and enhanced) and visible images

Daytime MTSAT-1R shortwave IR (original and enhanced) + MTSAT-1R visible images

A similar (but less obvious) image artifact could be seen on a night-time shortwave IR image over China (below) — there was a bright white “ghost”  to the  right of the warm area that was between the two colder cloud features (again, the ghost feature was offset to the east by 16 pixels with a vertical displacement of 1-2 pixels).

Night-time MTSAT-1R 3.75 µm shortwave IR image (over China)

Night-time MTSAT-1R 3.75 µm shortwave IR image (over China)

The exact cause of these image artifacts is not known; however, since the MTSAT-1R satellite scans from left to right using a Charge-Coupled Device (CCD) array,  the satellite sensor may be overcompensating for the CCD “quantum wells” losing more charge faster than expected, subtracting more of a bias than it should (this could also be a side-effect of sensor aging). Instrument cross-talk could be another source of this type of image anomaly.

Kudos to Chris Schmidt at CIMSS for processing and analyzing these MTSAT-1R images, and supplying the explanations of possible causes of such an image artifact.

Super Typhoon Choi-Wan (15W)

September 17th, 2009 |

MTSAT-1R IR images

MTSAT-1R IR images

MTSAT-1R 10.7 µm IR images (above) showed Super Typhoon Choi-Wan (17W) as it reached Category 5 intensity over the western North Pacific Ocean during the 15 September 16 September 2009 period. The eye of the typhoon was quite large (which moved just to the south of Pagan in the Northern Mariana Islands), with a very cold rings of convection seen during the various eyewall replacement cycles (as also seen on this MIMIC animation). A time series of the Advanced Dvorak Technique (ADT) intensity estimate from the CIMSS Tropical Cyclones site is shown below.

Advanced Dvorak Technique intensity estimates for Choi-Wan

Advanced Dvorak Technique intensity estimates for Choi-Wan

MODIS IR image

Terra MODIS IR image (14 September)

A pair of 1-km resolution MODIS images show the dramatic change in satellite appearance of Choi-Wan, from one with a ragged eye surrounded by cloud tops as cold as -87º C (darker purple color enhancement) on 14 September (above) to more of a classic  annular structure with a large and well-defined eye on 15 September (below).

MODIS IR image

Aqua MODIS IR image (15 September)

POES AVHRR and MODIS imagery in AWIPS

September 15th, 2009 |

GOES IR imagery

GOES IR images

AWIPS images of the GOES 10.7 µm IR window channel (above) showed that a patch of high-level clouds was moving southeastward over a band of low-level clouds that was in place across much of Ontario and Quebec, Canada on 15 September 2009. The GOES IR brightness temperatures associated with this patch of high clouds were in the -20º C to -40º C range (cyan to green color enhancement) — but the 4-km resolution of the GOES data over eastern Canada was significantly degraded during some of the pre-dawn hours due to the fact that GOES-12 (GOES-East) data was unavailable during that satellite’s Fall eclipse period.

While the 1-km resolution MODIS 11.0 µm IR window channel images (below) offered a more detailed depiction of the structure of the high cloud feature,  the temporal resolution of the MODIS images was far less than those available from GOES.

MODIS IR images

MODIS IR images

CIMSS has recently begun to test the experimental display of  NOAA POES Advanced Very High Resolution Radiometer (AVHRR) images and derived products in AWIPS (similar to the MODIS in AWIPS project initiated in 2006). One such derived AVHRR product is the 1-km resolution Cloud Type (below), which provides an indication of the phase of any particular cloud feature (e.g., liquid water droplets vs. ice crystals). The patch of high clouds moving southeastward across Ontario and Quebec was identified as a mixture of opaque ice (red), cirrus (yellow), and multi-layer (orange) — while the low cloud deck below it was identified as a mixture of fog (dark blue), liquid water droplet (medium blue), and supercooled liquid (cyan). This AVHRR Cloud Type product is one of a suite of products available from the CLAVR-x processing system.

AVHRR Cloud Type product

AVHRR Cloud Type product

As a demonstration of the accuracy of some of these MODIS and AVHRR derived cloud products, let’s now examine the cirrus cloud and contrail features seen later in the day across southern Wisconsin. These high cloud features were rather thin, as  evident on the comparison of the MODIS visible channel and the 1.3 µm near-IR “cirrus detection channel” images (below).

MODIS visible and cirrus channel images

MODIS visible and cirrus detection channel images

Since these high cloud features were so thin, a significant amount of radiation from the warm surface below was contaminating the IR brightness temperatures sensed by the satellite — note that the MODIS 11.0 µm IR values (below) were only as cold as -5º C to -15º C over the southern Wisconsin high cloud features, which would place the cirrus cloud and contrail features near the 500 hPa pressure level (19,000 – 22,000 feet above ground level) according to the morning rawinsonde data from Davenport, Iowa.

MODIS 11.0 µm IR image

MODIS 11.0 µm IR image + cloud height sampling skew-T

However, the 4-km resolution MODIS Cloud Top Temperature product (below) did suggest that some portions of these cirrus cloud and contrail features had temperatures as cold as -43º C (cyan color enhancement), which is closer to the expected temperature at the height of most cirrus clouds.

MODIS Cloud Top Temperature product

MODIS Cloud Top Temperature product

About an hour earlier, the 1-km resolution AVHRR Cloud Top Temperature values (below) were in the -45º C to -48º C range (cyan color enhancement), again much closer to the expected range of temperatures for such high-altitude  cloud features.

AVHRR Cloud Top Temperature product

AVHRR Cloud Top Temperature product

The AVHRR Cloud Top Height product (below) indicated that portions of these cirrus cloud and contrail features were as high as 10-11 km above the surface (orange color enhancement) — this was in good agreement with  ground-based lidar data from the SSEC Lidar Group, which showed the tops of the thicker cirrus clouds to be around 11 km.

AVHRR Cloud Top Height product

AVHRR Cloud Top Height product

GOES-14 sounder Derived Product Imagery

September 10th, 2009 |

GOES-14 sounder visible image

GOES-14 sounder visible image

GOES-14 (launched on 27 June 2009) is currently undergoing its Post Launch Test during the Summer and Fall of 2009. A GOES-14 sounder visible image (above) displayed  a variety of cloud systems across  the eastern US at 17:46 UTC on 10 September 2009. A preliminary version of the GOES-14 sounder Cloud Top Pressure derived product (below) indicated a wide range of cloud tops, from clusters of high (cold) clouds over the Gulf Coast region (blue to white color enhancement), to patches of low (warm) clouds over areas such as eastern Virginia/North Carolina and central Nebraska (orange to peach color enhancement).

GOES-14 sounder Cloud Top Pressure product

GOES-14 sounder Cloud Top Pressure product

Qualitatively, the GOES-14 sounder Cloud Top Pressure values (above) agreed well with those derived using GOES-12 sounder data (below). Note that both images are displayed using their respective native satellite projections.

GOES-12 sounder Cloud Top Pressure product

GOES-12 sounder Cloud Top Pressure product

Focusing on the aforementioned low cloud top features seen over eastern Virginia/North Carolina and over central Nebraska, both the GOES-12 and the GOES-14 Cloud Top Pressure (CTP) derived products indicated that those cloud tops were generally in the 700-800 mb range (orange color enhancement). Using the AWIPS cursor “Skew-T cloud height sampling” functionality, there was good agreement with the sounder CTP values: both the Moorhead City NC rawinsonde and the RUC model sounding over central Nebraska suggested that the GOES sounder IR brightness temperatures over those locations existed near the 700 mb pressure level (below).

GOES sounder IR image + Skew-T cloud height sampling

GOES sounder IR image + Skew-T cloud height sampling (rawinsonde)

GOES sounder IR image + Skew-T cloud height sampling

GOES sounder IR image + Skew-T cloud height sampling (RUC model)