Cloud Retrieval Using Infrared Sounder Data: Error Analysis

Bruce A. Wielicki National Center for Atmospheric Research, Boulder, CO 80307

Search for other papers by Bruce A. Wielicki in
Current site
Google Scholar
PubMed
Close
and
James A. Coakley Jr. National Center for Atmospheric Research, Boulder, CO 80307

Search for other papers by James A. Coakley Jr. in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

An error analysis is presented for cloud-top pressure and cloud-amount retrieval using infrared sounder data. Rms and bias errors are determined for instrument noise (typical of the HIRS-2 instrument on TIROS-N) and for uncertainties in the temperature profiles and water vapor profiles used to estimate clear-sky radiances. Errors are determined for a range of test cloud amounts (0.1–1.0) and cloud-top pressures (920–100 mb). Rms errors vary by an order of magnitude depending on the cloud height and cloud amount within the satellite's field of view. Large bias errors are found for low-altitude clouds. These bias errors are shown to result from physical constraints placed on retrieved cloud properties, i.e., cloud amounts between 0.0 and 1.0 and cloud-top pressures between the ground and tropopause levels. Middle-level and high-level clouds (above 3–4 km) are retrieved with low bias and rms errors. For instrument noise the 4.3 μm channels provide the smallest errors. For temperature profile and water vapor profile uncertainties the 15 μm channels generally give smaller errors. Errors due to rms temperature profile uncertainties of 2.0°C are found to be larger than errors due to instrument noise, independent of the sounding channels used.

Abstract

An error analysis is presented for cloud-top pressure and cloud-amount retrieval using infrared sounder data. Rms and bias errors are determined for instrument noise (typical of the HIRS-2 instrument on TIROS-N) and for uncertainties in the temperature profiles and water vapor profiles used to estimate clear-sky radiances. Errors are determined for a range of test cloud amounts (0.1–1.0) and cloud-top pressures (920–100 mb). Rms errors vary by an order of magnitude depending on the cloud height and cloud amount within the satellite's field of view. Large bias errors are found for low-altitude clouds. These bias errors are shown to result from physical constraints placed on retrieved cloud properties, i.e., cloud amounts between 0.0 and 1.0 and cloud-top pressures between the ground and tropopause levels. Middle-level and high-level clouds (above 3–4 km) are retrieved with low bias and rms errors. For instrument noise the 4.3 μm channels provide the smallest errors. For temperature profile and water vapor profile uncertainties the 15 μm channels generally give smaller errors. Errors due to rms temperature profile uncertainties of 2.0°C are found to be larger than errors due to instrument noise, independent of the sounding channels used.

Save