Uncertainties in space-based estimates of clouds and precipitation: implications for deriving global diabatic heating

Abstract

The Earth's weather and climate is driven by the exchange of energy between the sun, atmosphere, surface, and space and energy transport required to establish a global balance. Clouds and precipitation play an integral role in this exchange, enhancing reflection of solar radiation to space, trapping thermal emission from the surface, and providing a mechanism for the direct transfer of energy to the atmosphere through the release of latent heat in precipitation. As a result, there is an intimate coupling between the climate, energy budget, and global hydrologic cycle. The problem of establishing observational evidence for these connections and climate change in general, poses a significant challenge to the observational community. This dissertation seeks to address the components of this problem related to observing the hydrologic cycle and its role in modulating the tropical energy budget, from space-based measurements. This work reports on a new technique which makes use of cloud and precipitation information from the Tropical Rainfall Measuring Mission to estimate the principal components of the tropical energy budget and to examine the mechanisms by which clouds and precipitation modify it. First, three distinct retrieval algorithms are employed to determine the three-dimensional structure of cloud and precipitation in the tropical atmosphere. The first retrieves cloud and precipitation profiles from passive microwave observations from the TRMM Microwave Imager while the second applies a different technique to the same observations in an effort to derive estimates of non-precipitating liquid cloud. Finally, the third algorithm makes use of infrared radiances from the Visible and Infrared Scanner to infer ice cloud optical properties in non-precipitating regions. The resulting representation of the three-dimensional structure of cloud and precipitation in the tropical atmosphere is then used as input to a broadband radiative transfer model to derive profiles of short- and longwave fluxes. These flux profiles are composited to present a TRMM-based estimate of the short-term tropical energy budget for oceanic regions over the month of February 1998. On average, over this period, the tropical atmosphere absorbs 51 Wm-2 or 13 % of the 393 Wm-2 of solar radiation it receives. A further 112 Wm-2 is reflected by atmospheric particles, clouds, and the surface, leaving 230 Wm-2 to be absorbed by the ocean. At thermal wavelengths, it is found that the ocean emits 436 W m-2 of energy to the atmosphere while the atmosphere emits a total of 639 Wm-2 units, 407 Wm-2 downward toward the surface and 231 Wm-2 to space. Accounting for latent heat release which amounts to an exchange of 82 Wm-2 of energy between the surface and atmosphere, the results imply a deficit of 70 Wm-2 of energy in the atmosphere and a surplus of 121 Wm-2 at the Earth's surface. The implied net gain of 51 Wm-2 in the Earth-atmosphere system is consistent with a difference between the incoming solar radiation and emitted thermal radiation at the top of the atmosphere. It is speculated that these imbalances are largely accounted for by sensible heating, meridional energy transport, and absorption and transport of energy in the ocean. Finally, on average for the month of February 1998, the tropical atmosphere cools at -1 Kday-1 and experiences a net cloud forcing of -10 Wm-2 at TOA and -22 Wm-2 at the surface. A concerted effort has been made to rigorously characterize the uncertainties in all aspects of the approach. In the absence of additional tuning or constraints, the procedure described in the present work provides monthly-mean estimates of column radiative heating accurate to ~ 30 % and cloud radiative forcing with accuracies ranging from approximately 40 % for raining pixels to 75 % in non-precipitating clouds. It is shown that the dominant source of uncertainty in both the retrieval and radiative transfer models is a lack of vertical cloud boundary information inherent in the passive measurements. These results highlight the need for future algorithms to look toward making use of synergies between active and passive observations to simultaneously retrieve cloud and precipitation optical properties and their vertical distribution and ensure consistency between a wider variety of information sources. As a first step towards this undertaking, a new method for retrieving profiles of rainfall from spaceborne radars based on an optimal estimation technique is also introduced. The method is readily adapted to include information from a variety of sources and provides a suite of diagnostic tools with which to assess its performance. Preliminary results from synthetic retrievals highlight the utility of the algorithm for estimating profiles of precipitation up to 60 mmah-1 at 14 GHz and up to 8 mmh-1 at 94 GHz, provided some form of attenuation constraint is implemented. The technique described herein provides a complete approach for generating tropic-wide (and ultimately global) estimates of the components of the energy budget using explicit cloud and precipitation information from spaceborne observations. The results can, in principle, be applied to study short term climate variability through investigations of perturbations to the radiation balance induced by changes in the distributions of water vapor, cloud, and precipitation on short to moderate timescales affording us the opportunity to quantify important relationships between the hydrologic cycle and the Earth's energy budget.