@article{PAR00012387,
  title = {{G}lobal-scale comparison of passive ({SMOS}) and active ({ASCAT}) satellite based microwave soil moisture retrievals with soil moisture simulations ({MERRA}-{L}and)},
  author = {{A}l-{Y}aari, {A}. and {W}igneron, {J}. {P}. and {D}ucharne, {A}. and {K}err, {Y}ann and {W}agner, {W}. and {D}e {L}annoy, {G}. and {R}eichle, {R}. and {A}l {B}itar, {A}. and {D}origo, {W}. and {R}ichaume, {P}. and {M}ialon, {A}.},
  editor = {},
  language = {{ENG}},
  abstract = {{G}lobal surface soil moisture ({SSM}) datasets are being produced based on active and passive microwave satellite observations and simulations from land surface models ({LSM}). {T}his study investigates the consistency of two global satellite-based {SSM} datasets based on microwave remote sensing observations from the passive {S}oil {M}oisture and {O}cean {S}alinity ({SMOS}; {SMOSL}3 version 2.5) and the active {A}dvanced {S}catterometer ({ASCAT}; version {TU}-{W}ien-{WARP} 5.5) with respect to {LSM} {SSM} from the {MERRA}-{L}and data product. {T}he relationship between the global-scale {SSM} products was studied during the 2010-2012 period using (1) a time series statistics (considering both original {SSM} data and anomalies), (2) a space time analysis using {H}ovmoller diagrams, and (3) a triple collocation error model. {T}he {SMOSL}3 and {ASCAT} retrievals are consistent with the temporal dynamics of modeled {SSM} (correlation {R} > 0.70 for original {SSM}) in the transition zones between wet and dry climates, including the {S}ahel, the {I}ndian subcontinent, the {G}reat {P}lains of {N}orth {A}merica, eastern {A}ustralia, and southeastern {B}razil. {O}ver relatively dense vegetation covers, a better consistency with {MERRA}-{L}and was obtained with {ASCAT} than with {SMOSL}3. {H}owever, it was found that {ASCAT} retrievals exhibit negative correlation versus {MERRA}-{L}and in some arid regions (e.g., the {S}ahara and the {A}rabian {P}eninsula). {I}n terms of anomalies, {SMOSL}3 better captures the short term {SSM} variability of the reference dataset ({MERRA}-{L}and) than {ASCAT} over regions with limited radio frequency interference ({RFI}) effects (e.g., {N}orth {A}merica, {S}outh {A}merica, and {A}ustralia). {T}he seasonal and latitudinal variations of {SSM} are relatively similar for the three products, although the {MERRA}-{L}and {SSM} values are generally higher and their seasonal amplitude is much lower than for {SMOSL}3 and {ASCAT}. {B}oth {SMOSL}3 and {ASCAT} have relatively comparable triple collocation errors with similar spatial error patterns: (i) lowest errors in arid regions (e.g., {S}ahara and {A}rabian {P}eninsula), due to the very low natural variability of soil moisture in these areas, and {C}entral {A}merica, and (ii) highest errors over most of the vegetated regions (e.g., northern {A}ustralia, {I}ndia, central {A}sia, and {S}outh {A}merica). {H}owever, the {ASCAT} {SSM} product is prone to larger random errors in some regions (e.g., north-western {A}frica, {I}ran, and southern {S}outh {A}frica). {V}egetation density was found to be a key factor to interpret the consistency with {MERRA}-{L}and between the two remotely sensed products ({SMOSL}3 and {ASCAT}) which provides complementary information on {SSM}. {T}his study shows that both {SMOS} and {ASCAT} have thus a potential for data fusion into long-term data records.},
  keywords = {{SMOS} ; {ASCAT} ; {MERRA}-{L}and ; {S}oil {M}oisture ; {G}lobal {V}egetation ; {LAI}},
  booktitle = {},
  journal = {{R}emote {S}ensing of {E}nvironment},
  volume = {152},
  numero = {},
  pages = {614--626},
  ISSN = {0034-4257},
  year = {2014},
  DOI = {10.1016/j.rse.2014.07.013},
  URL = {https://www.documentation.ird.fr/hor/{PAR}00012387},
}