Cross Validation of Thermal Infrared Remotely Sensed Data In-Flight Using Automated Validation Sites © 2010 California Institute of Technology. Government sponsorship acknowledged. Dr Simon J. Hook William Johnson, Robert Radocinski, Glynn Hulley Jet Propulsion Laboratory, California Institute of Technology.

Outline Location and Characteristics –Lake Tahoe, CA/NV –Salton Sea, CA Measurements and Field Radiometer Calibration Data Reduction Methodology and Error Budgets Results from various sensors –MODIS, ASTER, Landsat Summary and Conclusions

Why Lake Tahoe? Large 35 km x 16 km High 2 km Available year round (does not freeze in winter). Homogenous compared with land. Large annual temperature range 5-25 C. Freshwater (kind to instruments!) Good infrastructure and easy access.

D

Measurements Offshore –bulk temperature, skin temperature, air temperature, wind speed, wind direction, relative humidity, net radiation. Onshore –air temperature, wind speed, wind direction, relative humidity, short and longwave radiation (up and down), sky imager, aerosols, total column water.

Site Layout and Measurement Stations

TB3 Installed Skin temperature Air temperature & Rel. Humidity Wind Speed & Direction Logging System Batteries Bulk Water Temperature 3m

8 Salton Sea, Platform Temperatures up to 35 C!

9 Radiometer Calibration and Cross-Cal

10 Cross Comparison of Radiometers at Miami JPL Rad

Data Reduction: Methodology For Radiance at Sensor Validation Extract the bulk temperatures. Extract the radiometric temperature. Correct the radiometric temperature to skin kinetic temperature. Propagate the skin temperature to the satellite using a radiative transfer model and interpolated atmospheric profile. Convolve the propagated at-sensor radiance to the instrument response function to obtain the Vicarious Radiance (VR). Extract the image radiance derived using the On Board calibrator (OBC). Compare and contrast the OBC and VR Radiance values.

Error Budgets PerturbationKinetic Temperature (degrees C)Difference from no perturbation (degrees C) None % Water Vapor % Water Vapor deg C deg C Ozone Factor 50% Ozone factor 150% um SRF shift um SRF shift Change from Nominal Temperature in Kelvin Channel 14 (center 8.55 um) Channel 15 (center ) Channel 16 (center ) 90% Water Vapor % Water Vapor deg C deg C % Ozone % Ozone % Visibility (11.5 km) % Visibility (34.5 km) um SRF shift um SRF shift

Several hundred matchups per year. MODIS OBC radiance agrees with Vicarious radiance in clear parts of spectrum

MODIS land and water bands show excellent agreement with vicarious radiance. R2 typically 0.99

Validation process works well for angles up to 30 degrees off nadir

10 years of validation! Values for each year from several hundred matchups. MODIS-Terra has performed within specification for duration of mission. Proof necessary for climate change studies

Validation measurements are made 24x7, validate mid infrared (3-5 um) at night to avoid reflected solar. Results shown for Aqua

Aqua MIR performance meets preflight specification

ASTER Short and Long-Term Calibration ASTER calibrates when instrument drifts. Introduces “sawtooth” appearance For much of mission predictable behavior but recently showed more rapid change

ASTER Short and Long-Term Calibration More frequent drift corrections now required and additional monitoring

If carefully monitor and correct can maintain instrument within specification, If look at mean radiance differences for each year they are typically less than 0.5% or 0.3K and 2006 need more examination.

22 High temperature points from Salton Sea allow evaluation of gain and offset. Landsat required gain and offset adjustment

After adjustment Landsat 7 meets specification

Summary and Conclusions Established an automated site for validating thermal infrared data at Lake Tahoe CA/NV. Site has been operating since Second site established at Salton Sea CA in Salton Sea provides very high temperatures. Measurements made at the site include skin- bulk- air- temperature, wind speed, wind direction and net radiation at multiple locations every 2 minutes. Multiple locations (4 buoys) allow validation of several points within a scene at Lake Tahoe. Validated data from multiple instruments including, ATSR2, AATSR, ASTER, MODIS (Terra, Aqua), Landsat 5 and Landsat ETM+, MTI. Results so far indicate: –MODIS-Terra and Aqua LWIR meet specification (no bias, abs. acc. ± 0.2K) –MODIS-Aqua MIR meet specification (Terra MIR does not) –ASTER, Landsat require adjustment but perform to specification In-flight validation is a necessity for climate studies!

Backup

Mirror AOI related to scan angle, validation process works well for angles up to 30 degrees off nadir

8 years of validation! Values for each year from several hundred matchups. MODIS-Aqua has performed within specification for duration of mission. Proof necessary for climate change studies

Skin Effect After Minnett et al Water surface

Similar issue with Landsat 5. After adjustment Landsat 5 meets specification. Landsat 5 data are slightly noisier than Landsat 7

37 Cross Comparison of Radiometers at Miami

38 Means and standard deviations of the estimated skin SST differences between pairs of radiometers for the entire cruise period, and for each half of the cruise The Third International Infrared Radiometer Workshop to be held in Miami in the late spring of 2009

39 Variations in the bulk-, skin-, air-temperatures and wind speed on June 7, 2001 at L. Tahoe

40 Land Surface Temperature Results MODIS, ASTER, AATSR

BACKUP