 |
NRL Monterey, Marine Meteorology Division
|
|---|
|
NRL Monterey, Marine Meteorology Division
|
|---|
| Rainrate example for Mitch | Track Map for Mitch discussed below |
|---|---|
) |
) |
| The SSM/I rainrate image gives reasonable estimates of precipitationrates over wide expanses of ocean. Although it is just a snapshotin time (called "instantaneous" rainrate) available just a few times aday, it is an important source of quantitative information. The rainratewithin tropical cyclones can give important clues about its intensityand radius of influence. The SSM/I rain rate gives a useful estimate ofinstantaneous rain rate falling at the surface of the earth. It iscalibrated in mm/hr. By instantaneous, it means that it is only valid ata particular instant in time. Were it to rain at this same rate, say 10mm/hr, for an hour on a particular spot on the earth, 10 mm of rain would have fallen. See below in Ferraro et al. for the rainrate algorithm weuse here. Over land, the algorithm uses a precipitation scatteringalgorithm only. This means that precipitation is estimated by using thebrightness temperatures (Tb) depression that results from precipitatingconvection. Over the ocean the algorithm uses a combinedscattering/emission algorithm. Emission is the microwave warming thatresults from raindrops. Since both emission and scattering signals areused over the oceans, the rainfall retrievals over oceans are moreaccurate than over land. Over coastlines, it is impossible to retrieveprecipitation because the microwave pixels represent a combination ofocean and land pixels, precluding the application of an algorithm. TheSSM/I algorithm gives precipitation rate at 25 km. Reference: Ferraro, R.R., 1997: Special sensor microwave imager derivedglobal rainfall estimates for climatological applications. J. Geophys.Res., 102, D14, 16715-16735. | |
| Link to Paul McCrone's Tutorial on SSM/I, AFWA/Metsat Applications:Tutorial |
| Other than the data from isolated buoys, this rainrate is the mostaccurate measurement of rainrate available over most of the oceans of theworld. It benefits from the ability of the SSM/I to "see through" cloudtops to observe the underlying precipitation structure. It is thereforeunlike infrared precipitation techniques that estimate precipitationamount based on cloud top temperature. It can show, for example,asymmetries within the precipitation structure of a tropical cyclone. Although images of 85 GHz can show storm structure at a higher resolutionthan the rainrate product (12 km vs. 25 km), interpretation of 85 GHzimages is complicated by the ambiguous signatures of emission, scatteringand sea surface emissivity. The rainrate image enables a much morestraightforward interpretation. By viewing a series of images, theanalyst can often infer storm intensification. When the area of rainingpixels increases, or the precipitation amounts increase, a storm isprobably intensifying. When the area of rain pixels decreases, or theprecipitation amounts decrease, a storm is probably weakening. |
| 1. The rainrate product is not computed near coastlines; thus, noinformation is available about rainrate at storm landfall. 2. Caution should be used for rainrates over land, especially inmountainous terrain. Over land, the algorithm is sensitive to microwavescattering from convection, yielding reasonable precipitation estimates. However, the algorithm is insensitive to microwave emission from raindropsover land. Therefore, the algorithm will not show areas of precipitationwhich exist without an ice phase. 3. Since the resolution of the rainrate product is 25 km, it does thecapture the detail seen in images of 85 GHz or its derivatives, such asthe PCT or the color composite. For example, even in intense storms, eyesare often not apparent in the rainrate images. 4. Very light rainrates or drizzle (~ 1 mm/hr or less) are often notrepresented well in the rain rate images. Frequently, very light rainfallrates are classified as "no rain" or visa versa by the rainratealgorithm. Although the absolute errors in rainrate are not great in thisinstance (< 1 mm/hr), the fact that the existence of precipitation is notdetermined correctly negatively impacts various applications. To theextent that an application is sensitive to very light rainrates (e.g., theestimation of surface visibility in drizzle), caution should be takenbefore using rainrate as a yes/no precipitation flag.5. The 25-km size of an individual rainrate retrieval is an average ofprecipitation rates within that pixel. There may be places within thatpixel with much higher rates and places with much lower rates. |
| Example I: Rainrate over Mitch | Example I: 85 GHz Color Composite |
|---|---|
) |
) |
| This tutorial presents four examples of the rainrate for HurricaneMitch in October 1998. This storm dropped torrential rain overCentral America, much of which was observed by SSM/I sensors. On theright of the following four rainrate examples is an 85 GHz colorcomposite image, which shows storm convection in red. Low cloudbands over water and land appear as blue-green.To the left above, the first SSM/I color composite shows Mitchapproaching Central America. The hurricane circulation and eyeare readily apparent. Away from the storm over the west coastof Central America, areas of red indicate heavy convective showers.The rainrate image (left) shows quantitative rainrates for manyareas especially offshore. The heaviest rates near the center ofthe storm are near 20 mm/hr (0.80 inches/hr). Blues away from thecenter of the storm show rates at 0-10 mm/hr (0 - 0.40 inches/hr). At 25 km resolution (as opposed to 12 km for the color compositeon the right) the rainrate product is not as easy to interpret as animage. It is difficult to identify with certainty the banded andcellular nature of the precipitation. A major drawback is the "absence"of precipitation over coastlines in the rainrate product, for example,over the northern coast of Honduras. This may lead the viewer tobelieve that it is not raining in this region. However, the colorcomposite image (left) suggests that it is raining heavily in thisregion. | |
| Example II: Rainrate over Mitch | Example II: 85 GHz Color Composite |
|---|---|
) |
) |
| The 85 GHz color composite image (right) shows the stormcenter just north of the coast of Honduras. Intense spiralbands (red) dominate the image. The rainrate image (left) showsquantitative precipitation rates over the same area. Blues are thelightest rates; greens indicate heavier precipitation; reds are theheaviest. Note that the entire northern coast of Honduras and theentire eastern coast of Nicaragua shows no precipitation. Precipitationis actually falling in this region, but the rainrate algorithm can notdetect it, hence the strip of white along the coasts. The same effectappears on the west coast of Central America. The color compositeimage suggests precipitation (red); however, the rainrate algorithmleaves a white strip along the coastline. | |
| Example III: Rainrate over Mitch | Example III: 85 GHz Color Composite |
|---|---|
) |
) |
| The storm center has moved onshore in Honduras as shown by the colorcomposite image on the right. This image shows what appears to be an eyeembedded within a spiral band structure. Dark red over western Nicaraguasuggests heavy precipitation. The rainrate image (left) showsprecipitation rates over this same area. It captures the heavyprecipitation that flooded Nicaragua during the period but shows white (noprecipitation) along the west coast. This effect is because the algorithmcannot detect the precipitation there. | |
| Example IV: Rainrate over Mitch | Example IV: 85 GHz Color Composite |
|---|---|
) |
) |
| The final pair of images from this series shows Mitch continuing to affectNicaragua, Honduras, and Guatemala with torrential precipitation. Thecomposite image (right) shows a large, nearly-rain free storm center overwestern Honduras (region free of red). The rainrate image (left) capturesthe large precipitation rates over the region, especially over centralHonduras and southwest of Honduras. However, it fails to capture thedetail shown by the composite product (right). Because it cannot detectprecipitation near coastlines, it does not capture the high rates on theborder of Nicaragua and Honduras on the west coast. The viewer should notbelieve that no precipitation is occurring in this region. | |
| GOES Infrared Color Image | GOES Visible Image corresponding |
|---|---|
) |
) |
| The final set of images above are the infrared (left) and visible (right)images corresponding to the passive microwave images in the previoustwo-panel. These images show that heavy precipitation as seen in thepassive microwave images generally corresponds to cold cloud tops inthe infrared and bright clouds in the visible. | |
Author: 7541tl\@nrlmry.navy.mil Last Updated: Mon Nov 25 16:48:12 2002 Produced by: The Composer (Ver: 1.09 ) |
|