Use of a GCM to explore sampling issues in connection with satellite remote sensing of the Earth radiation budget

摘要

Collocated in time and space, top-of-the-atmosphere measurements of the Earth radiation budget (ERB) and cloudiness from passive scanning radiometers, and lidar- and radar-in-space measurements of multilayered cloud systems, are the required combination to improve our understanding of the role of clouds and radiation in climate. Experiments to fly multiple satellites "in formation" to measure simultaneously the radiative and optical properties of overlapping cloud systems are being designed. Because satellites carrying ERB experiments and satellites carrying lidars- or radars-in space have different orbital characteristics, the number of simultaneous measurements of radiation and clouds is reduced relative to the number of measurements made by each satellite independently. Monthly averaged coincident observations of radiation and cloudiness are biased when compared against more frequently sampled observations due, in particular, to the undersampling of their diurnal cycle. Using the Colorado State University General Circulation Model (CSU GCM), the goal of this study is to measure the impact of using simultaneous observations from the Earth Observing System (EOS) platform and companion satellites flying lidars or radars on monthly averaged diagnostics of longwave radiation, cloudiness, and its cloud optical properties. To do so, the hourly varying geographical distributions of coincident locations between the afternoon EOS (EOS-PM) orbit and the orbit of the ICESAT satellite set to fly at the altitude of 600 km, and between the EOS PM orbit and the orbits of the PICASSO satellite proposed to fly at the altitudes of 485 km (PICA485) or 705 km (PICA705), are simulated in the CSU GCM for a 60-month time period starting at the idealistic July 1, 2001, launch date. Monthly averaged diagnostics of the top-of-the-atmosphere, atmospheric, and surface longwave radiation budgets and clouds accumulated over grid boxes corresponding to satellite overpasses are compared against monthly averaged diagnostics obtained from hourly samplings over the entire globe. Results show that differences between irregularly (satellite) and regularly (true) sampled diagnostics of the longwave net radiative budgets are the greatest at the surface and the smallest in the atmosphere and at the top-of-the-atmosphere, under both cloud-free and cloudy conditions. In contrast, differences between the satellite and the true diagnostics of the longwave cloud radiative forcings are the largest in the atmosphere and at the top-of-the atmosphere, and the smallest at the surface. A poorer diurnal sampling of the surface temperature in the satellite simulations relative to the true simulation contributes a major part to sampling biases in the longwave net radiative budgets, while a poorer diurnal sampling of cloudiness and its optical properties directly affects diagnostics of the longwave cloud radiative forcings. A factor of 8 difference in the number of satellite overpasses between PICA705 and PICA485 and ICESAT leads to a systematic factor of 3 difference in the spatial standard deviations of all radiative and cloudiness diagnostics. [References: 40]