resolutions sufficient to resolve the diurnal cycle) of less than 30 W m⁻² at 10-km resolution and over the open ocean with an accuracy of 5 W m^–2 at a spatial resolution of 1° (about 100 km). Those errors, although still substantial, would be small enough to be comparable with those in other terms in the global (and regional) water and energy budgets and so would facilitate direct estimation of evaporation errors rather than estimation as a residual term, as is often done (see Roads et al., 2003, for an example of water and energy budget estimation for the Mississippi River basin).

Under turbulent conditions, evaporation is directly proportional to latent heat flux (a component of the surface energy balance) and carbon flux in the surface carbon balance. Evaporation also codetermines—with precipitation—the rate of the global water cycle. The difference between evaporation and precipitation should be zero when globally aggregated, which gives a long-range performance goal for this grand challenge.

The difficulty in measuring evaporation from space is that bulk parameterizations, which are the primary means for estimating evaporation, require knowledge of the near-surface specific humidity, a measurement that continues to elude the scientific community. Even space-borne profilers with very high vertical resolution are unable to resolve the boundary layer with the needed precision. Other quantities necessary for estimation of the latent heat flux are surface wind speed and surface and near-surface temperatures. Over the oceans, surface wind estimates are possible with both active and passive microwave instruments with reasonable accuracy, although under high wind conditions only scatterometers have proven utility. Over land, direct measurement of surface wind from space is not now possible. Although not a direct input to the latent heat flux parameterization, surface temperature is needed to determine saturation vapor pressure at the surface in the case of ocean evaporation, and the actual surface humidity in the case of evaporation over land. Furthermore, the surface and air temperatures together determine the stability of the surface layer, which affects the transfer coefficients used in the calculation of latent heat flux.

Sea-surface and land temperature measurements with a variety of current and planned sensors in both the visible-infrared and microwave wavelengths can provide diurnally varying values with a relatively high level of accuracy, and a continued mix of space-borne microwave radiometers will continue this record. It should be noted that in addition to the satellite-data limitations there are still unresolved issues with the bulk flux parameterizations themselves (Curry et al., 2004).

Remote sensing of land radiometric surface temperature (LST) is critical for all current schemes to estimate evapotranspiration remotely. LST is directly related to the sensible heat component of the energy balance and is thus inversely proportional to latent energy and evaporation rates. The Bowen ratio (H/LE) summarizes the relationship between sensible and latent heat flux from a surface. Thermal remote sensing can provide an integrated look at land surface evaporation, although overpass timing is critical (midafternoon radiant heating of the land surface provides the most useful signal). For some purposes, data from the Geostationary Operational Environmental Satellites (GOES) also can be used to derive LST and surface evapotranspiration every hour under cloud-free conditions.

Other methods for inferring evaporation can, with a combination of measured and modeled techniques, give some understanding of this flux over large areas. For instance, atmospheric budget analysis using moisture convergence in combination with observed precipitation can be used to estimate evaporation by difference—a technique that is applicable over both land and ocean. Over the oceans, changes in upper-ocean salinity combined with oceanic advection can be used to produce an estimate of E—P (global time-varying salinity measurements from Aquarius are expected to improve the basis for estimating space-time fields of E—P over the oceans). In both cases, knowledge of precipitation is necessary—a constraint that is especially limiting over the oceans and portions of the land where precipitation is poorly observed. Other promising techniques involve the fusion of satellite data with global or regional climate-

Front Matter (R1-R26)

Executive Summary (1-16)

Part I: An Integrated Strategy for Earth Science and Applications from Space, 1 Earth Science: Scientific Discovery and Societal Applications (17-26)

2 The Next Decade of Earth Observations from Space (27-60)

3 From Satellite Observations to Earth Information (61-78)

Part II: Mission Summaries, 4 Summaries of Recommended Missions (79-140)

Part III: Reports from the Decadal Survey Panels, 5 Earth Science Applications and Societal Benefits (141-151)

6 Human Health and Security (152-189)

7 Land-Use Change, Ecosystem Dynamics, and Biodiversity (190-216)

8 Solid-Earth Hazards, Natural Resources, and Dynamics (217-256)

9 Climate Variability and Change (257-303)

10 Weather Science and Applications (304-337)

11 Water Resources and the Global Hydrologic Cycle (338-380)

Appendix A Statement of Task (381-384)

Appendix B Biographical Information for Committee Members and Staff (385-391)

Appendix C Blending Earth Observations and Models -- The Successful Paradigm of Weather Forecasting (392-409)

Appendix D Request for Information from Community (410-412)

Appendix E List of Responses to Request for Information (413-422)

Appendix F Acronyms and Abbreviations (423-428)