1. Main Issues:

Validating two- or three- parameters retrievals from multi-angular observations over a variety of scenes and vegetation types

• Correcting for the effects of vegetation:
• Characterizing the dependence of optical depth t and w as a function of view angle, polarization, and vegetation type.
• Accounting for within-pixel heterogeneity (mixed pixel include open water surfaces, forests, agricultural areas, ...)
• forest signature at L-band
• snow at L-band
• special events (frozen soils, dew effects, rain interception ...)
• Correcting for the effects of topography (invariant with time) and surface roughness
• Improving knowledge of surface roughness effects, are they significant at SMOS scale?
• Testing methods to account for topography
• Testing methods on actual 2-D interferometric observations (accounting for actual sensor capabilities and performances)

Questions and actions:

• Radiometers and data sets exist already: Any new action should capitalise on these so as to optimize scientific outputs.
• for instruments : coordinate actions on required instrument characteristics vs what exists
• for data sets: pool existing measurements and design future campaigns accordingly (i.e., types of vegetation not well covered, adequate instrumentation, ...)
• campaigns: coordinate actions over different vegetation types

• Pool all existing data sets over soil and vegetation surfaces and document them.
• Coordination of future experiments (campaign design , measurement protocols, ...)
• Define standard calibration procedures

2. Experiments which are most needed:

(1) for the vegetation cover types where SM retrieval is feasible (grass, crops, ...), it is necessary to improve the retrieval algorithm over this specific vegetation cover, by accounting for the dependence of optical depth on polarization and incidence angle.

(2) for some other vegetation canopy types (dense forests, ..),  SM retrievals amy be questionable, but these canopies will be included in SMOS pixels. For instance, forest canopies will be included in most SMOS pixels in Europe.
Thus, an accurate characterization of the microwave signatures of forests in required in order to perform SM retrievals from time variations of the average pixel emissivity (in relation with time variations of emissivity in non-forested areas).

Five main types of vegetation cover categories were experiments are most needed, were defined:
 

	% Coverage	Scheduled Experiments & existing or near future Study
Arid / Semi Arid areas	30%	FAUGA Toulouse (J-C Calvet) long term Ground Based measurements (end 2001 - 2003 )  A. Van de Griend / L. Simmonds (Gr. based , 2001 ?)
Grass lands	20%
Agricultural crops mainly:  wheat  corn  sunflower  soybean		Avignon 2001, (JP Wigneron , INRA / TUD) 4 -month, Gr. based, on corn, wheat, bare soils polarimetric measurements  Univ of Reading (L Simmonds) 2001. 'multiple angle measurements over crops'  Orgeval site (C. Loumagne, C Ottley), airborne meas. (2002?), wheat  Studies:   INRA/CETP : (Wigneron, Pardé, ..): 'On-going study based on existing round-based data sets (wheat, corn, soybean, ...)'
Forests		Valencia site (Spain) E. Lopez airborne measurements (2001 ?)  Northern forests (M Hallikainen, HUT) airborne measurements (2001 ?)  Studies:   Univ. Roma (P. Ferrazzoli and L Guerriero) : 'On-going simulation study based on discrete modeling and ground-based data sets'
Tundra areas

 

3. Calibration / Validation site

• Valencia site (agricultural, matorral, mediterranean forests) (E. Lopez, Univ. of Valencia) .At the present time, this site is the main site considered for SMOS validation studies in the near future. Other sites are:
• Africa (Botswana, site used in previous studies of A. Van de Griend et al.)
• Northern Scandinavia
• Upper Danube

4. Dielectric constant

• Measurements of both soil and vegetation dielectric constants (?) are needed

5. Within-pixel heterogeneity

1) ESA ITT  : AO/1-3652/00/NL/DC, 'Study on soil moisture retrieval by a future space-borne earth observation mission'. (2 years - accepted Nov 2000), Resp. L Simmonds (Univ of Reading) in relation will several european scientists. Task 2 and 3 concerning direct modeling and retrieval studies will investigate the microwave emission from SMOS pixels (at spatial scale ~ 40 km) and will account for within-pixel heterogeneity (using land cover maps at scale of ~ 5 km)

In complement of the simulation studies, airborne measurements at different flight altitudes could be carried out over heterogenous scenes (including agricultural areas , grassland and forests ...) in the S/W of France.

2) NASA Proposal, 'Remotely sensed soil moisture from SMOS in land data assimilation systems with heterogeneous land cover', J. Shuttleworth and E. Burke (Univ of Arizona)
Develop and test methods that will allow patch specific estimates of soil moisture in the NASA-Goddard and NCEP Land data assimilation system (LDAS) from multiple measurements of pixel-average TB data.

Other studies are to be carried out in Europe:

• A. Vand de Griend & L Simmonds : simulation study of within-pixel heterogeneity effects, by combining microwave signatures of several crop fields.
• JP Wigneron & M Pardé (Alpilles site): simulation over an actual agricultural site (Alpilles) during one year, by aggregation of TB for each field, obtained by coupling the INRA crop growth model (STICS) with TB models.
• M Hallikainen, Scandinavia site (HUT): airborne measurements at different flight altitudes could be carried out to study heterogenous scenes (including frozen soils, snow, forests, ...). Funding will be required for this airborne campaign.

6. Topography

• It seems that based on existing data sets and expertise, it is possible to correct for this effect, since,
•  topography characteristics are invariant with time (in general !)
•  Digital elevation Model (DEM) can provide accurate information on topography at rather fine resolution

• We do not know well what is the effect of topography on the accuracy of the soil moisture retrievals.

7. Roughness effects

•  Probably, simple modeling approach including both topography and small scale roughness effects can be developed for the SMOS retrieval algorithm.
• Roughness effects are mostly invariant, except over agricultural areas. Over such areas, seasonal change in the soil roughness characteristics, in relation with the agricultural practices and rainfall events, may have an impact on the land surface emissivity.