For further information, refer to validation reports and collected papers on TES validation.
The most recent version string for all TES data products is F04_04. TES Level 1B radiances (current version is V002) have undergone rigorous validation. TES Level 2 products (current version is V003) have undergone extensive quality control and validation testing, and include updates to improve nadir temperature products and all limb products.
Currently, the TES L2 products that are that are validated with known biases and ready for scientific use are nadir retrievals of ozone, carbon monoxide, temperature and water vapor.
||Validated Stage 2
|Nadir Carbon Monoxide
||Validated Stage 2
||Validated Stage 2
||Validated Stage 2
|Sea Surface Temperature
||Validated Stage 2
|Land Surface Temperature & Emissivity
||Validated Stage 1
|Limb Nitric Acid
|Limb Water Vapor
||Beta (Validation Ongoing)
Note: TES L2 retrievals include fully characterized internal error estimates and do not obtain error estimates from external sources. Uncertainty in the TES validation work describes biases when compared to other data sources. Refer to the validation status definitions for more information on validation terms used here.
Current validation needs and requirements
A variety of field campaigns, during which in situ and remote sensing measurements are made by instruments on aircraft and balloons, contribute to the validation of TES products. TES validation reports and related publications are used to document these efforts. Special observations can usually be scheduled to coincide with field campaigns. Data inter-comparisons are made between TES and in situ data, and between TES and other satellite instruments, including MOPITT, MLS, OMI and AIRS. TES data is also examined in conjunction with models of global-scale air quality and climate. Prior validation campaigns in which TES has participated are listed at the bottom of this page. Data needed for ongoing validation is listed in the table below.
||more “clear sky” data for monitoring L1B data quality
||better characterization of sea ice in Arctic and Antarctic
||needed between 500-700 hPa
||mid-trop profile data, ARCTAS and TC4-Guam
|TATM, CH4, H2O, O3
||ozone and CFH sondes
||CFH sondes needed for clear sky and ocean scenes to identify systematic errors
||improve statistics, ARC-IONS, tropical ocean scenes, TC4-Guam
Species by species status
In order to investigate systematic error sources, TES validation efforts include examination of spectral residuals using the additional information available in the spectrum itself, radiance closure studies and joint retrievals of atmospheric temperature with water vapor, ozone, surface temperature, and cloud. Consistency is critical to identifying potential sources of systematic error, since systematic errors in one will propagate into the others.
TES ozone (O3) profiles have been compared to ozonesonde and lidar measurements. These comparisons show that TES generally sees higher O3 in the lower and middle troposphere than the sondes and lidar. The magnitude of this difference varies somewhat with different geographic regions. In the Southern low and middle latitudes, in the uppermost troposphere, TES sees lower values than the sondes and lidar.
In addition, TES measurements of total, stratospheric and tropospheric column abundance of O3 are being compared with data from other satellite instruments (including [[MLS]] and [[OMI]]) in order to better quantify how the different sensitivities of TES and the other instruments affect the observed instrument biases as a function of season and latitude. Comparisons of TES total column O3 with [[OMI]] (Ozone Monitoring Instrument) show similar global distributions, but TES measures 3-7% more O3. The source of these biases is under investigation. Clouds and aerosols affect the vertical sensitivity of TES, and this information is captured by the averaging kernel (provided with each retrieval).
Nadir Carbon Monoxide
TES retrievals of carbon monoxide (CO) were much improved after the optical bench warm up in early December 2005 as a result of the better alignment of the instrument and increased signal to noise (a four-fold increase in signal-to-noise ratio at higher frequencies).
Comparisons have been carried out between TES CO retrievals and those from a variety of satellite and aircraft instruments. Global patterns of CO as measured by TES are in good qualitative agreement with those seen by MOPITT (Measurement Of Pollution In The Troposphere). Comparisons of profiles of CO between TES and MOPITT show good agreement when 'a priori' information is accounted for correctly. Enhanced CO data in the lower troposphere can be closely related to the known burning or pollution sources. TES CO profiles are compared to all satellite CO data from currently operating instruments -- MOPITT, AIRS, ACE and MLS. These comparisons show general agreement in patterns of CO global distributions in the troposphere. TES CO also agrees to within the estimated uncertainty of aircraft instruments, including both errors and the variability of CO itself. In general, the agreement is better for regions where CO fields have less variability. Further validation of TES CO is being accomplished with use of the MOZAIC data set, which includes a variety of cities in different regions for an extended term time period.
Validation for TES water vapor (H2O) is undertaken through a combination of lidar, sonde, and aircraft measurements in the upper troposphere, and through statistical comparisons between TES, AIRS, and sondes. TES V003 water vapor shows a 5% improvement over V002 below 500 hPa. Comparisons to [[CFH]] (Cryogenic Frostpoint Hygrometer) indicate that TES is 5-10% wetter below 700 hPa, and 5-40% wetter at 300 to 700 hPa. Conclusions from comparisons of TES water retrievals with sondes and aircraft data are difficult due to atmospheric variability, though radiance closure experiments suggest that differences in the middle and upper troposphere cannot be fully accounted for by known systematic errors.
The main differences between V002 and V003 that influence the TES water vapor retrievals are: (i) improved TES temperature retrievals due to inclusion of the CO2 spectral region with improved CO2 forward model calculations (Shephard et al., 2007b); (ii) the migration of TES initial guess and a priori from GEOS-4 to GEOS-5; (iii) a lowered minimum value for the a priori cloud optical depth in order to better handle clouds with lower optical depths; and (iv) the addition of more surface microwindows to help characterize the surface.
The temperature field must be as accurate as possible in order for good retrievals of atmospheric chemical species to be made. TES temperature retrievals have been compared extensively with both remote sensing and 'in situ' measurements. The V003 retrieval uses additional microwindows in the CO2 band at 650 to 800 cm-1. V003 temperature bias is improved relative to V002, except for a 0.5 to 2 K cold bias at 400-500 hPa, and a warm bias at 800 hPa.
Sea Surface Temperature
TES measurements of sea surface temperature from V003 have been compared against TES V002 for eight global surveys. The comparison shows a mean difference of 0.2 K with a standard deviation of 2.3 K. Comparisons demonstrate radiometric stability, and retrievals under clear sky conditions (effective cloud optical depth of ~0.05) are not biased by a priori surface temperature.
Version 2 TES data uses an improved Level 1B calibration algorithm that brings TES into very good agreement with the aircraft instrument Scanning High-Resolution Interferometer Sounder S-HIS and the Atmospheric Infrared Sounder AIRS instrument. TES radiances show mean differences of less than 0.3 K at brightness temperatures of 290-295 K, and less than 0.5 K at brightness temperatures of 265-270 K with both Scanning HIS and AIRS.
Land Surface Temperature and Emissivity
Land surface emissivity is more difficult to characterize than ocean emissivity. Calibration of all TES Level 1 products can benefit from coincident cloud top pressures over uniform, thin clouds, as well as from cloudfree coincidences.
Initial comparisons between TES, DACOM (Differential-Absorption Carbon Monoxide Monitor) (DC-8) and ground-based FTIR (Fourier Transform Infrared Spectrometer) indicate that TES methane is biased ~5% high in column measurements, particularly between 150 and 500 hPa pressure levels.
TES estimates of deuterium (HDO) have undergone preliminary validation by comparison with models and aircraft data. A bias of approximately 5% has been seen, but the distribution of HDO/H2O as measured by TES and the JPL instrument ALIAS (Aircraft Laser Infrared Absorption Spectrometer) shows good agreement.
Limb Nitric Acid
Comparisons have been carried out between TES, Microwave Limb Sounder MLS Chemical Ionization Mass Spectrometers (CIMS) (WB-57 aircraft), and Soluble Acidic Gases and Aerosol (SAGA) (on the DC-8 aircraft) during INTEX-B. TES has low sensitivity to nitric acid in the troposphere due to clouds. TES stratospheric nitric acid shows similar spatial distributions to these other measurements.
TES limb temperature retrievals have undergone preliminary validation by comparison with [[GMAO]] (Global Modeling Assimilation Office) GEOS-5. In the troposphere, TES limb temperature has a bias of -0.08 K relative to GMAO. In the stratosphere, TES limb temperature typically has a warm bias that gradually increases with altitude, up to +2 K at 12 hPa.
TES limb water vapor retrievals have undergone preliminary validation by comparison with GMAO GEOS-5. TES limb retrievals have sensitivity to water in the middle troposphere (biased ~4% low relative to GMAO) but little sensitivity at other altitudes.
Go to the field campaigns
page to view TES science browsing plots for previous validation campaigns and field experiments.
TES validation takes place by comparing quantities measured by TES with other measurements that represent the state of the atmosphere. Validation efforts may use a combination of satellite observations, in situ measurements, and chemistry and transport models, each of which represent the state of the atmosphere in different ways. Validation helps scientists quantify systematic errors, and can provide constraints for atmospheric models into which TES data are assimilated.