98
R.O. Dubayah, E.F. Wood, E.T. Engman et al.
between point observations and areal estimates. One promising approach is the use
of data assimilation methodologies that merge observations with models. For example, TIROS Operational Vertical Sounder (TOVS) observations are used with a
GCM in a data assimilation which then predicts fields of air temperature at coarse
spatial resolution.
Lastly, the exploitation of distributed forcing fields generated from remotelysensed data requires an evolution in hydrologic model structure and a rethinking of
how such data are best used. Current structures are parameterized based on those
data that are most frequently measured on the ground. This places a burden on
remote sensing science by requiring derivation of variables such as air temperature
that are difficult to estimate from space, while ignoring those that are easier, such
as surface temperature. One possibility is to use satellite-measured skin temperature to update model skin temperatures, and thus keep the model from diverging
from reality over time. The same method could be taken with other variables such
as near-surface soil moisture state or snow cover, as measured with passive or
active microwave sensors such as SSMR and SSMII.
Other approaches may scrap parameterizations that involve several components,
each of which have large uncertainties as derived from satellites, in favor of ones
that have fewer elements. An example of the latter is the parameterization of latent
heat flux in terms of net radiation. It has been shown that net radiation may be
estimated empirically using solar radiation, which in turn is found relatively accurately from remote sensing. Whether or not this is an appropriate avenue of exploration for macroscale modeling remains to be decided. What is clear is that the
capabilities of current and future remote sensing instruments, such as the EOS
suite of sensors, must be studied carefully relative to hydrological modeling, and
the models redesigned to fully take advantage of the data these instruments produce.
Acknowledgments
Thanks to Nancy Casey-McCabe for providing research assistance. This work is
supported by NASA Grant NAGW-5l94 (Dubayah) and NASA Contract NAS531719 (Wood).
References
Abdulla, F. A., D.P. Lettenmaier, E. F. Wood and J.A. Smith, Application of a Macroscale
Hydrological Model to Estimate the Water Balance of the Arkansas-Red River Basin, J. Geophys. Res., 1996
Allord, GJ. and Scarpace, F.L. Improving Streamflow Estimates Through Use of Landsat. In
Satellite Hydrology, 5th Annual William T. Pecora Memorial Symposium on Remote Sensing, Sioux Falls, SD. pp 284-291, 1979
Agbu, P.A. B. Vollmer and M.E. James, Pathfinder AVHRR Land Data Set. NASA Goddard
Space Flight Center, Greenbelt. 1993
Arola, A., D.P. Lettenmaier, and E.F. Wood, Some preliminary results of GCIP modeling activities in the Arkansas-Red River basin, First International Scientific Conference on the Global
Energy and Water Cycle Royal Society, London, 1994
R.O. Dubayah, E.F. Wood, E.T. Engman et al.
between point observations and areal estimates. One promising approach is the use
of data assimilation methodologies that merge observations with models. For example, TIROS Operational Vertical Sounder (TOVS) observations are used with a
GCM in a data assimilation which then predicts fields of air temperature at coarse
spatial resolution.
Lastly, the exploitation of distributed forcing fields generated from remotelysensed data requires an evolution in hydrologic model structure and a rethinking of
how such data are best used. Current structures are parameterized based on those
data that are most frequently measured on the ground. This places a burden on
remote sensing science by requiring derivation of variables such as air temperature
that are difficult to estimate from space, while ignoring those that are easier, such
as surface temperature. One possibility is to use satellite-measured skin temperature to update model skin temperatures, and thus keep the model from diverging
from reality over time. The same method could be taken with other variables such
as near-surface soil moisture state or snow cover, as measured with passive or
active microwave sensors such as SSMR and SSMII.
Other approaches may scrap parameterizations that involve several components,
each of which have large uncertainties as derived from satellites, in favor of ones
that have fewer elements. An example of the latter is the parameterization of latent
heat flux in terms of net radiation. It has been shown that net radiation may be
estimated empirically using solar radiation, which in turn is found relatively accurately from remote sensing. Whether or not this is an appropriate avenue of exploration for macroscale modeling remains to be decided. What is clear is that the
capabilities of current and future remote sensing instruments, such as the EOS
suite of sensors, must be studied carefully relative to hydrological modeling, and
the models redesigned to fully take advantage of the data these instruments produce.
Acknowledgments
Thanks to Nancy Casey-McCabe for providing research assistance. This work is
supported by NASA Grant NAGW-5l94 (Dubayah) and NASA Contract NAS531719 (Wood).
References
Abdulla, F. A., D.P. Lettenmaier, E. F. Wood and J.A. Smith, Application of a Macroscale
Hydrological Model to Estimate the Water Balance of the Arkansas-Red River Basin, J. Geophys. Res., 1996
Allord, GJ. and Scarpace, F.L. Improving Streamflow Estimates Through Use of Landsat. In
Satellite Hydrology, 5th Annual William T. Pecora Memorial Symposium on Remote Sensing, Sioux Falls, SD. pp 284-291, 1979
Agbu, P.A. B. Vollmer and M.E. James, Pathfinder AVHRR Land Data Set. NASA Goddard
Space Flight Center, Greenbelt. 1993
Arola, A., D.P. Lettenmaier, and E.F. Wood, Some preliminary results of GCIP modeling activities in the Arkansas-Red River basin, First International Scientific Conference on the Global
Energy and Water Cycle Royal Society, London, 1994