ESCOMPTE/CLU

Brigthness surface temperature images
from NOAA Advanced Very High Resolution Radiometers
June 4th -July 14th 2001

Benedicte Dousset
Hawaii Institute of Geophysics and Planetology,
& Geosystemes, UMR 6554, IUEM/UBO. Brest.

1. Satellites and instruments

The NOAA satellites are launched to an altitude of about 820 km in near-polar sun-synchronous orbits. Since 3 satellites are active simultaneously, data were available 6 times a day, in between:  ~01:10 - 02:40, and ~12:10 - 13:40 UTC for NOAA 16; ~04:20 - 05:50 and 15:20 - 16:30 UTC for NOAA 14; and ~04:50 - 05:40 and 15:40 - 16:40 UTC for NOAA12.

The Advanced Very High Resolution Radiometer (AVHRR) on board these satellites is a scanning radiometer with five spectral channels, centered in the visible and near-infrared at 0.62 µm, (channel 1) and 0.91 µm (channel 2), and in the thermal infrared at 3.74 µm (channel 3), 10.8 µm (channel 4) and 12.0 µm (channel 5). The measurement noise is ~0.1 oC in these last two channels. Note that NOAA 16 has 6 channels (including channels 3a and 3b). The scan angle is 112 oC, resulting in a swath width on the earth of 2970 km with a ground resolution (pixel size) of 1.1 x 1.1 km2 at the sub-satellite point degrading to 1.5 x 4.0 km2 at the swath edge.

The AVHRR data were received at the High Resolution Picture Transmission (HRPT) station of Modena in Italy (F. Parmigiani), and subsequently processed at the University of Hawaii Satellite Oceanography Laboratory, including calibration, navigation and remapping to a common grid. A series of 159 images were processed covering a domaine of 120 x 120 km centered at 43o 30.00 N and 5o 20.00 E.   More information on the NOAA polar orbiting satellites can be found in Kidwell (1995).

2. Data acquisition and processing

The raw HRPT telemetry was first archived onto 4 mm data tapes. Images were processed with a combination of Terascan software, and locally generated routines, along the following stages.

The first stage extracts the AVHRR spectral channels from the HRPT records, in the geographic domain of interest. Earth location was computed using satellite orbit elements supplied daily by the U.S. Department of  Defense. Ancillary data was also extracted from the tape, to be subsequently used for radiometrically calibrating the AVHRR data, that is, converting the raw digital counts to blackbody equivalent brightness temperature in oC (for channels 3, 4, and 5).

The second stage of processing involved a visual display of the extracted AVHRR data on an interactive monitor. Many readily identifiable land points along the coastlines were used to correct the spacecraft clock time, and the pointing angle of the AVHRR scanner. Interactive routines were used to bring the land points seen in the image into alignment with the land points obtained from a digital data base of the coastlines. This was done by first adjusting the spacecraft clock time, which is inaccurate by up to 1 second, translating into a 6 pixel location error along the satellite track. Following the time correction, another correction was made for inaccuracies in the attitude control system of the spacecraft. The pitch, roll, and yaw stabilization is good to 0.2 degrees, which translates to about two pixels in the AVHRR. The final absolute ground location accuracy is estimated to be 2 km.

The third stage of processing draws upon the ancillary calibration information to produce radiometrically calibrated data in % albedo and oC, following Kidwell (1995). Prior to scanning the earth, the AVHRR first views cold space at 3oK. Following the earth scan, the instrument views the inside of its housing, within which are embedded four platinum resistance thermometers. From this and prelaunch calibration constants, one may convert the raw count data to radiance units. The radiances are then converted to blackbody equivalent temperatures.

The fourth stage of processing consists of remapping the data to a common geographical grid, a process called registration. The geometric distortion and earth location - line/sample relation is different from one satellite overpass to the next. This is corrected by establishing a fixed grid, which is then populated by resampling the initial data set on a pixel by pixel basis. The simplest nearest neighbor algorithm is used. The result is a geometrically fixed set of final output pixels of 1 km x 1 km size, allowing quantitative comparisons between images taken at different times.  Finally,  from a data set of more than 300 images, 159 images were selected according to satellite-zenith angles comprised between 3o8 and 51o9, to allow the study of directionnal effects.

Temperatures obtained at this point are Brightness Temperatures as seen by the AVHRR viewing through the earth's atmosphere. Brightness temperatures are typically colder than actual surface temperature for a variety of reasons. Most of these, namely clouds, water vapor, aerosols, and departure of the surface from being a true blackbody, tend to lower the brightness temperature by an amount ranging from one to several oC, depending upon atmospheric conditions. To derive Land Surface Temperatures (LST) from Brightness Temperatures it is necessary to use a radiative transfer model, and near-simultaneous radiosounding recorded during the ESCOMPTE experiment. Over the ocean (a near-perfect blackbody), an empirical multi-spectral correction for water vapor is generally computed based on the differential attenuation of infrared channels 4 and 5 (McClain et al., 1985). This correction does not apply to land surfaces given the spatial and spectral variation of their emissivity.

Clouds have not been masked for the data base, since their observation and statistics are important in studying surface temperature processes.  They may be masked as a second iteration when needed. Darzi (1992) reviews methods from various authors for detecting and screening cloud contaminated pixels for polar orbiting satellite radiometers and discusses criteria for comparing and evaluating methods.  Clouds may be screen by textural test, based on the variations in temperature due to the uneven cloud-top heights of most convective clouds, which are detectable when comparing channel 4 brightness temperatures between neighboring pixels. The non-uniform nature of most cloud tops also causes variations in reflectance received by the visible and near-infrared channels in daytime. Threshold determination can be done in the same manner as for channel 4, with tests on daytime images only.  Stratiform clouds can be detected through the differences between the infrared channels. The three infrared wavelengths have different sensitivities to emitted surface radiation, emitted atmospheric radiation, reflected solar radiation, and absorption by water vapor; in particular, clouds have different optical properties at different wavelengths (Saunders and Kriebel, 1988). The emissivity of dense water clouds (ie, low stratus or fog) is less for channel 3 than for channels 4 and 5 (0.7 vs. 0.99), but it is greater for ice clouds and high thin cirrostratus clouds. Cirrus and less dense low relatively warm clouds are more transparent to channel 3 than to channels 4 or 5. For these clouds the satellite radiometer partially picks up surface radiation or penetrates deeper into the cloud, receiving radiation emitted at lower and warmer levels. Clear surface seen through a moist atmosphere will also give channel 3 values greater than channel 4, since absorption by water vapor is larger for channel 4. Bi-variate histograms of the three pairs of infrared channels can be constructed and analyzed to determine which difference thresholds between the channels would successfully flag the various cloud types found in the domain.

3. List of images

In the following list, the first colum gives the satellite number (NOAA 12, 14 or 16), the year (01), the julian day (from 155 to 195), and the hour in UTC. The second column gives the Sun-Zenith angle, the third column gives the Satellite-Zenith angle and the fourth column gives the Sun-Reflection angle. The solar zenith angle (Sun-Zenith) is the angle between the sun and a line perpendicular to the Earth's surface at the view point.  The satellite zenith angle (Sat-Zenith), is the angle  between the satellite and a line perpendicular to the Earth's surface at the view point.  The solar reflection angle (Sun-Reflection) is the angle at the view point between vectors pointing toward the satellite and the sun's specular reflection.

        Sat.Y Day. UTC 	Sun-Zen.        Sat-Zen.	Sun Ref.
	________________________________________________________

	n12.01155.1554	57.59553	9.202623	48.7288
	n14.01156.0541	72.89074	43.72304	40.55119
	n16.01156.1241	25.22722        16.49435        15.43277
	n14.01156.1528	52.70292        37.06339        18.77005
	n16.01157.0243	100.6714        50.29453        78.68957
	n12.01157.0525	75.55477        29.63704        51.88151
	n16.01157.1230	23.92500        29.78617        18.12914
	n14.01157.1516	50.43257        47.92044        11.67863
	n16.01158.0232	101.9117        41.78313        83.44271
	n12.01158.0502	79.50457        4.879479        82.76646
	n14.01158.0517	76.98297        15.99898        63.92974
	n16.01158.1220	22.76999        40.71634        27.51022
	n12.01158.1625	62.83952        35.04839        96.97674
	n16.01159.0222	103.1048        31.09751        89.42948
	n12.01159.0439	83.36775        35.92897        113.3561
	n14.01159.0505	79.00498        4.012993        80.56973
	n16.01159.1209	21.78459        49.44114        36.83363
	n12.01159.1601	58.51575        4.265917        60.65119
	n14.01159.1633	64.22208         44.8451        107.8338
	n16.01160.0211	104.2477        18.03679        96.54372
	n14.01160.0453	81.00735         19.6129        97.41641
	n12.01160.1538	54.19582         31.2598        24.93938
	n14.01160.1621	61.95089        32.38831        93.44458
	n16.01161.0201	105.3376        4.368952        104.0676
	n14.01161.0441	82.9872         34.86773        111.8946
	n12.01161.0533	74.20023        37.83568        45.12363
	n14.01161.1609	59.67938        16.75549        75.94292
	n16.01162.0150	106.3713        12.34536         110.662
	n14.01162.0429	84.94164        46.87427        123.0951
	n12.01162.0509	78.18565        7.472579        72.04893
	n16.01162.1319	29.88962         35.6101        63.67022
	n14.01162.1557	57.40985        3.816005        56.40957
	n12.01162.1633	63.79359        42.84466        105.5013
	n16.01163.0140	107.3458        26.56552         115.158
	n12.01163.0446	82.08967        27.00827        104.6952
	n16.01163.1309	28.24209        23.02921        49.69783
	n14.01163.1544	55.14493         18.5383        37.49878
	n12.01163.1609	59.4882         13.48044         72.5836
	n16.01164.0129	108.2579        38.57328        117.3456
	n16.01164.1258	26.68126        8.413963        34.24503
	n14.01164.1532	52.88734        33.47128        22.01959
	n12.01164.1545	55.18019        21.85896        34.39189
	n16.01165.0119	109.1047        48.25742        117.7723
	n14.01165.0534	73.97435        35.33421        46.90726
	n16.01165.1247	25.22319        7.079165        19.94283
	n14.01165.1520	50.64015        45.19344        12.84538
	n12.01166.0518	76.94888        18.12939        62.22655
	n16.01166.1237	23.88655        21.55571        14.51282
	n16.01167.0239	100.8006        47.50807        80.50569
	n12.01167.0454	80.89376        16.92341        95.05272
	n14.01167.0509	78.11894         4.50938         75.4084
	n16.01167.1226	22.69245        34.00385        21.53195
	n12.01167.1616	60.51606        23.97036        83.83326
	n16.01168.0229	102.0661        38.23941        85.59884
	n12.01168.0430	84.74319        44.72583         121.301
	n14.01168.0457	80.16441        14.79972        92.50085
	n16.01168.1216	21.66385        44.09788        31.44328
	n12.01168.1552	56.22311        11.31993        45.35537
	n14.01168.1626	62.2457         36.31183        97.51254
	n16.01169.0218	103.284         26.65713        91.93774
	n14.01169.0445	82.18867        30.92612        107.8252
	n12.01169.0547	71.73396        50.62872        36.62858
	n12.01169.1529	51.94028        40.69672        15.65605
	n14.01169.1613	60.00569        21.52039         80.8828
	n16.01170.0207	104.4515        12.74877        99.23309
	n14.01170.0433	84.18893        43.83748        119.9664
	n12.01170.0524	75.79008          27.791        53.80689
	n16.01170.1337	32.08106        51.20049        81.35365
	n14.01170.1601	57.76658        4.885332        61.71271
	n16.01171.0157	105.5654        4.215455        106.4992
	n12.01171.0500	79.77546        6.252606        84.78252
	n16.01171.1326	30.36512        42.31819        70.79149
	n14.01171.1549	55.53166        13.84256        42.31606
	n12.01171.1623	61.60356        33.25036        93.98011
	n16.01172.0146	106.6231        17.86995        112.3364
	n12.01172.0437	83.67172        37.35518        114.4588
	n16.01172.1316	28.71955        31.23467        58.17953
	n14.01172.1537	53.30331        29.67361        25.66154
	n12.01172.1600	57.32748         3.94046        57.23971
	n16.01173.0136	107.6211        31.38174         115.888
	n14.01173.0540	73.32666        38.82753        44.20804
	n16.01173.1305	27.15778        17.82138        43.55851
	n12.01173.1536	53.05631        32.88897        22.47723
	n16.01174.0125	108.5564        42.53095        117.3005
	n14.01174.0526	75.44875        25.12871        55.53713
	n16.01174.1254	25.69523        4.216631        28.12887
	n14.01174.1513	48.8792        	51.97971         12.4322
	n16.01175.0115	109.4257        51.44557        117.2316
	n16.01176.1233	23.14265         26.4494	16.48517
	n12.01176.1607	58.49744        11.12505        69.31347
	n14.01176.1631	62.78895        39.92912        101.5766
	n16.01177.0235	101.5196        44.40117        82.65776
	n16.01177.1223	22.09539        38.02702        25.69012
	n12.01177.1543	54.24032        23.82543        31.59675
	n14.01177.1618	60.58342        26.01149        85.83993
	n16.01178.0225	102.8154        34.28479        88.22098
	n14.01178.0438	83.76177        40.58496        117.0501
	n12.01178.0538	73.68703        43.31686        41.82234
	n16.01178.1212	21.23157        47.31708        35.40501
	n12.01178.1520	50.00394        48.41993        11.26483
	n14.01178.1606	58.37844        9.100824        67.19352
	n16.01179.0214	104.0634        21.75517        94.99972
	n14.01179.0426	85.78349        51.26886        126.9459
	n12.01179.0514	77.75258        15.92052        64.84331
	n14.01179.1554	56.17698        9.029332        47.51861
	n12.01179.1638	63.97394        47.93285        110.7149
	n16.01180.0203	105.2606        7.136043        102.4515
	n12.01180.0451	81.7413         19.06248        97.60053
	n14.01180.0555	70.79405        51.78975        35.76003
	n16.01180.1333	31.25804        48.12439        77.32658
	n14.01180.1542	53.98207        25.67178        29.79462
	n12.01180.1614	59.73751         21.6829        80.87444
	n16.01181.0153	106.404         8.536752          109.33	
	n14.01181.0543	72.96883        41.99679        41.94873
	n16.01181.1322	29.6346         38.45893        66.08284
	n14.01181.1530	51.79712        39.22132        16.66071
	n12.01181.1550	55.4968         13.61302        42.39668
	n16.01182.0144	107.4905        23.30855        114.3047
	n14.01182.0531	75.13202        29.19842         52.1997
	n12.01182.0545	72.73192         49.3658        38.02086
	n16.01182.1311	28.09328         26.4768        52.74016
	n14.01183.0519	77.28141        13.16095        66.54742
	n16.01183.1301	26.64881        12.28281          37.684
	n16.01184.0121	109.4801        46.25767        117.5971
	n12.01184.0458	80.87007        8.415378        87.86869
	n14.01184.1635	63.61549        43.25099        105.6898
	n14.01185.0455	81.53026        22.39471         100.226
	n16.01185.1240	24.12068        18.07954          15.139
	n12.01185.1558	56.82989        4.356635        54.28891
	n14.01185.1623	61.447        	30.20832        90.83945
	n16.01186.1229	23.07713        31.15111        20.21283
	n14.01186.1611	59.27899        13.97439        72.8558
	n16.01187.0231	102.7999        40.92999        85.20753
	n12.01187.0529	75.97499        34.38355        49.30996
	n14.01187.1559	57.11496        4.923649        53.13226
	n16.01188.0221	104.131        	29.88579        91.37261
	n12.01188.0505	80.05325        4.339609        77.85413
	n16.01188.1208	21.53719        50.38263        39.17259
	n14.01188.1546	54.95774        21.46359        34.50172
	n12.01188.1628	62.44693        39.30167         100.907
	n16.01189.0210	105.414         16.41822        98.64257
	n14.01189.0548	72.8624         44.87815        39.95943
	n14.01189.1534	52.81109        35.93218        19.66954
	n16.01190.0159	106.646         3.847136        106.1656
	n14.01190.0535	75.0657         32.95423        49.26562
	n14.01190.1523	50.67879        47.16316         11.5825
	n16.01191.0149	107.824         14.41117        112.4919
	n14.01191.0523	77.25697        17.76176         62.7545
	n12.01191.0536	75.16351        41.78655        43.69036	
	n16.01191.1318	29.50496        34.24773        61.49573
	n12.01191.1518	49.88467        49.54075        10.01145
	n16.01192.1307	28.09641        21.34732        47.51866
	n12.01193.0449	83.33198        21.41389        101.2108
	n16.01193.1257	26.79831        6.597721        32.41118
	n12.01193.1612	59.74097        19.17924        78.53542
	n14.01193.1628	62.62847        34.11176        95.91685
	n14.01194.1615	60.50051	18.65895	78.69863
	n12.01195.0543	74.39621	48.07918	39.35102
	n14.01195.1603	58.37653	 3.82698	59.14485

4. Access to the data and thumbnails

Individual matlab format images, and thumbnails of the NIR albedo and TIR brightness temperature, are accessible by clicking here.

Alternately, right-click here to download a compressed tar file containing all the individual matlab images.

A simple matlab m-file to display and annotate images is given here.


5. Bibliography

1. Darzi, M., "Cloud screening for polar orbiting visible and infrared (IR) satellite sensors, "SeaWiFS Technical Report Series, Volume 7, NASA Technical Memorandum 104566, pp. 1-7, 1992.
2. Kidwell, K.B., "NOAA Polar Orbiter Data Users Guide," 1995.
3. McClain, E.P., W. Pichel, and C. Walton, "Comparative performance ofAVHRR-based multichannel sea surface temperatures," J. Geophys. Res., vol. 90, pp. 11587-11601, 1985.
4. Saunders, R.W. and K.T. Kriebel, "An improved method for detecting clear sky and cloudy radiances from AVHRR data," Int. J. Remote Sensing, vol. 9, pp. 123-150, 1988.