The 27–28 October 1986 FIRE IFO Cirrus Case Study: Cloud Parameter Fields Derived from Satellite Data

View More View Less
  • 1 Atmospheric Sciences Division, NASA Langley Research Center, Hampton, Virginia
  • | 2 Lockheed Engineering and Sciences Company, Hampton, Virginia
© Get Permissions
Full access

Abstract

A methodology for estimating cirrus cloud amounts and altitudes using visible and infrared satellite data was developed and tested using FIRE Cirrus Intensive Field Observation (IFO) coincident lidar and satellite data with a theoretical cloud albedo model. On average, cloud center heights could be determined to within ±0.9 km of the lidar-derived values using the satellite data alone. Satellite-derived, total cloud tops are generally 0.5 ± 0.9 km lower than the lidar cloud tops. If only high clouds are considered, the avenge cloud top is 0.1 ± 0.6 km higher than the lidar cloud top. The accuracies of the lidar cloud-center and cloud-top heights are estimated to be within ±0.7 km of the actual values. Satellite-derived average cloud emittance and visible optical depths can be determined to within ±0.05 and ±0.13, respectively, of the reference cloud emittance. Cirrus cloud thickness was also derived. The satellite retrieval yields cloud depths that are 0.3±1.0 km thinner than the lidar-derived cloud thicknesses. The accuracy of the lidar-derived cloud depths is estimated to be 0.7 km. It was concluded that compared to a method which analyzes each pixel individually, a bispectral approach, which averages some of the pixel values before analysis, yields lower rms and bias errors in some of the derived parameter values.

The technique was applied to GOES and AVHRR data taken during the daylight hours of the FIRE Cirrus IFO case study on a 0.5° grid covering most of Wisconsin. Broadband radiation fields from the ERBE corresponding to the AVHRR results were also analyzed. During the afternoon of 27 October 1996, a cirrus field was tracked with the GOES data as it developed over northern Wisconsin. The satellite analyses revealed average cloud-top heights ranged between 9 and 11 km. Decreases in the outgoing longwave fluxes caused by the cloud appeared to be balanced by increases in the cloud albedo resulting in a negligible change in energy balance at the top of the atmosphere due to the cloud. During 28 October, the cloud fields were highly variable with both cirrus and midlevel clouds. An organized cirrus “wedge” developed and passed though the region during the middle of the day with cloud-top heights greater than 11 km. In addition to other cirrus clouds, an apparent cirrus convective complex passed through central Wisconsin during the afternoon with cloud tops between 10.0 km and the tropopause at ∼11.3 km. A north–south line of clearing with scattered altocumulus separated the morning and afternoon cloud fields. This paper provides a comprehensive, quantified description of the case study clouds and should be useful for verifying ISCCP results and for improving the understanding of cirrus processes when combined with other IFO measurements.

Abstract

A methodology for estimating cirrus cloud amounts and altitudes using visible and infrared satellite data was developed and tested using FIRE Cirrus Intensive Field Observation (IFO) coincident lidar and satellite data with a theoretical cloud albedo model. On average, cloud center heights could be determined to within ±0.9 km of the lidar-derived values using the satellite data alone. Satellite-derived, total cloud tops are generally 0.5 ± 0.9 km lower than the lidar cloud tops. If only high clouds are considered, the avenge cloud top is 0.1 ± 0.6 km higher than the lidar cloud top. The accuracies of the lidar cloud-center and cloud-top heights are estimated to be within ±0.7 km of the actual values. Satellite-derived average cloud emittance and visible optical depths can be determined to within ±0.05 and ±0.13, respectively, of the reference cloud emittance. Cirrus cloud thickness was also derived. The satellite retrieval yields cloud depths that are 0.3±1.0 km thinner than the lidar-derived cloud thicknesses. The accuracy of the lidar-derived cloud depths is estimated to be 0.7 km. It was concluded that compared to a method which analyzes each pixel individually, a bispectral approach, which averages some of the pixel values before analysis, yields lower rms and bias errors in some of the derived parameter values.

The technique was applied to GOES and AVHRR data taken during the daylight hours of the FIRE Cirrus IFO case study on a 0.5° grid covering most of Wisconsin. Broadband radiation fields from the ERBE corresponding to the AVHRR results were also analyzed. During the afternoon of 27 October 1996, a cirrus field was tracked with the GOES data as it developed over northern Wisconsin. The satellite analyses revealed average cloud-top heights ranged between 9 and 11 km. Decreases in the outgoing longwave fluxes caused by the cloud appeared to be balanced by increases in the cloud albedo resulting in a negligible change in energy balance at the top of the atmosphere due to the cloud. During 28 October, the cloud fields were highly variable with both cirrus and midlevel clouds. An organized cirrus “wedge” developed and passed though the region during the middle of the day with cloud-top heights greater than 11 km. In addition to other cirrus clouds, an apparent cirrus convective complex passed through central Wisconsin during the afternoon with cloud tops between 10.0 km and the tropopause at ∼11.3 km. A north–south line of clearing with scattered altocumulus separated the morning and afternoon cloud fields. This paper provides a comprehensive, quantified description of the case study clouds and should be useful for verifying ISCCP results and for improving the understanding of cirrus processes when combined with other IFO measurements.

Save