Abstract
Recent advances in L-band passive microwave remote sensing provide an unprecedented opportunity to monitor soil moisture at ~40 km spatial resolution around the globe. Nevertheless, retrieval of the accurate high spatial resolution soil moisture maps that are required to satisfy hydro-meteorological and agricultural applications remains a challenge. Currently, a variety of downscaling, otherwise known as disaggregation techniques have been proposed as the solution to disaggregate the coarse passive microwave soil moisture into high-to-medium resolutions. These techniques take advantage of the strengths of both the passive microwave observations of soil moisture having low spatial resolution and the spatially detailed information on land surface features that either influence or represent soil moisture variability. However, such techniques have typically been developed and tested individually under differing weather and climate conditions, meaning that there is no clear guidance on which technique performs the best. Consequently, this paper presents a quantitative assessment of the existing radar-, optical-, radiometer-, and oversampling-based downscaling techniques using a singular extensive data set collected specifically for that purpose, being the Soil Moisture Active Passive Experiment (SMAPEx)-4 and -5 airborne field campaigns, and the OzNet in situ stations, to determine the relative strengths and weaknesses of their performances. The oversampling-based soil moisture product best captured the temporal and spatial variability of the reference soil moisture overall, though the radar-based products had a better temporal agreement with airborne soil moisture during the short SMAPEx-4 period. Moreover, the difference between temporal analysis of products against in situ and airborne soil moisture reference data sets pointed to the fact that relying on in situ measurements alone is not appropriate for validation of spatially enhanced soil moisture maps.
Original language | English |
---|---|
Article number | 111586 |
Pages (from-to) | 1-23 |
Number of pages | 23 |
Journal | Remote Sensing of Environment |
Volume | 239 |
Early online date | 14 Jan 2020 |
DOIs | |
Publication status | Published - 15 Mar 2020 |
Funding
This research was made possible through the financial support from an ARC Discovery Project (MoistureMonitor, DP140100572 ) and ARC Linkage Infrastructure, Equipment and Facilities grant ( LE0453434 ). Monash University is also acknowledged for its contribution towards a postgraduate scholarship for Sabah Sabaghy to pursue her PhD. Jian Peng was supported by the ESA SEOM project “Exploitation of S-1 for Surface Soil Moisture Retrieval at High Resolution” under Contract 4000118762/16/I-NB . Appendix A Fig. A1 Land cover maps showing dominant vegetation cover at 1 and 9 km spatial resolution the same as that of downscaled soil moisture maps. Fig. A1 Table A1 Summary table on the relative accuracy [-] of soil moisture downscaling products at 1 km derived from their temporal analysis against OzNet in situ and airborne PLMR soil moisture estimates. Table A1 Against OzNet Against airborne PLMR (SMAPEx-4 and -5) Against airborne PLMR (SMAPEx-4) Against airborne PLMR (SMAPEx-5) Downscaling technique Downscaled product Bias RMSD ubRMSD Bias RMSD ubRMSD Bias RMSD ubRMSD Bias RMSD ubRMSD [-] [-] [-] [-] [-] [-] [-] [-] [-] [-] [-] [-] SMOS DisPATChA 0.113 0.415 0.390 −0.204 0.633 0.510 −0.273 0.743 0.587 −0.083 0.401 0.170 Optical-based SMOS DisPATChD 0.060 0.527 0.516 −0.099 0.538 0.451 −0.160 0.604 0.468 −0.134 0.488 0.329 SMAP VTCI 0.080 0.387 0.305 −0.294 0.731 0.604 − − − −0.189 0.488 0.303 SMOS VTCI 0.005 0.540 0.494 −0.203 0.493 0.341 − − − −0.421 0.519 0.230 SMAP PassiveA 0.110 0.395 0.363 −0.300 0.691 0.588 −0.318 0.648 0.549 −0.272 0.482 0.271 Uniform field SMAP PassiveD 0.244 0.410 0.321 −0.222 0.594 0.496 −0.148 0.394 0.292 −0.233 0.486 0.282 SMOS PassiveA 0.141 0.336 0.314 −0.231 0.675 0.598 −0.189 0.643 0.567 −0.246 0.510 0.330 SMOS PassiveD 0.180 0.566 0.507 −0.240 0.648 0.563 −0.099 0.481 0.379 −0.297 0.548 0.381 Table A2 As for Table A1 but for products at 9 km. Table A2 Against OzNet Against airborne PLMR (SMAPEx-4 and -5) Against airborne PLMR (SMAPEx-4) Against airborne PLMR (SMAPEx-5) Downscaling technique Downscaled product Bias RMSD ubRMSD Bias RMSD ubRMSD Bias RMSD ubRMSD Bias RMSD ubRMSD [-] [-] [-] [-] [-] [-] [-] [-] [-] [-] [-] [-] Radar-based SMAP MOEA 0.311 0.492 0.381 −0.111 0.330 0.271 −0.111 0.330 0.271 − − − SMAP A/P 0.404 0.770 0.646 −0.242 0.488 0.341 −0.242 0.488 0.341 − − − SMOS DisPATChA −0.072 0.293 0.284 −0.241 0.612 0.558 −0.367 0.724 0.627 −0.148 0.336 0.260 Optical-based SMOS DisPATChD −0.085 0.433 0.424 −0.195 0.543 0.487 −0.169 0.531 0.428 −0.216 0.412 0.310 SMAP VTCI −0.148 0.241 0.191 −0.335 0.654 0.568 − − − −0.266 0.424 0.304 SMOS VTCI −0.145 0.459 0.439 −0.238 0.327 0.222 − − − −0.423 0.465 0.178 Radiometer-based SMAP SFIM 0.209 0.372 0.311 −0.178 0.486 0.432 −0.043 0.358 0.288 −0.209 0.378 0.319 Oversampling-based SMAP EnhancedA 0.020 0.295 0.294 −0.254 0.537 0.465 −0.098 0.288 0.223 −0.294 0.376 0.225 SMAP EnhancedD 0.093 0.289 0.274 −0.233 0.494 0.433 −0.181 0.324 0.274 −0.264 0.379 0.255 SMAP PassiveA 0.070 0.331 0.315 −0.322 0.611 0.516 −0.356 0.647 0.546 −0.337 0.405 0.229 Uniform field SMAP PassiveD 0.177 0.322 0.275 −0.249 0.505 0.423 −0.183 0.326 0.252 −0.301 0.395 0.250 SMOS PassiveA 0.068 0.269 0.260 −0.240 0.594 0.534 −0.196 0.587 0.551 −0.251 0.422 0.303 SMOS PassiveD 0.102 0.477 0.466 −0.250 0.546 0.513 −0.101 0.395 0.361 −0.301 0.456 0.307 Table A3 Summary table on the relative accuracy [-] of soil moisture downscaling products at 1 km derived from their spatial analysis against airborne PLMR soil moisture maps. Table A3 Against airborne PLMR (SMAPEx-4 and -5) Against airborne PLMR (SMAPEx-4) Against airborne PLMR (SMAPEx-5) Downscaling technique Downscaled product Bias RMSD ubRMSD Bias RMSD ubRMSD Bias RMSD ubRMSD [-] [-] [-] [-] [-] [-] [-] [-] [-] SMOS DisPATChA 0.009 0.643 0.537 −0.085 0.833 0.548 −0.077 0.524 0.520 Optical-based SMOS DisPATChD −0.082 0.579 0.544 −0.082 0.680 0.561 −0.216 0.584 0.501 SMAP VTCI −0.127 0.520 0.446 − − − −0.132 0.554 0.484 SMOS VTCI −0.266 0.572 0.495 − − − −0.421 0.645 0.495 SMAP PassiveA −0.068 0.522 0.465 −0.086 0.479 0.451 −0.231 0.559 0.517 Uniform field SMAP PassiveD −0.128 0.499 0.455 −0.181 0.456 0.422 −0.200 0.549 0.519 SMOS PassiveA 0.007 0.512 0.469 0.121 0.476 0.461 −0.175 0.547 0.527 SMOS PassiveD −0.180 0.545 0.501 −0.122 0.540 0.438 −0.393 0.615 0.533 Table A4 As for Table A3 but for products at 9 km. Table A4 Against airborne PLMR (SMAPEx-4 and -5) Against airborne PLMR (SMAPEx-4) Against airborne PLMR (SMAPEx-5) Downscaling technique Downscaled product Bias RMSD ubRMSD Bias RMSD ubRMSD Bias RMSD ubRMSD [-] [-] [-] [-] [-] [-] [-] [-] [-] Radar-based SMAP MOEA −0.090 0.351 0.273 −0.090 0.351 0.273 − − − SMAP A/P −0.087 0.525 0.509 −0.087 0.525 0.509 − − − SMOS DisPATChA −0.042 0.447 0.327 −0.086 0.544 0.393 −0.138 0.347 0.306 Optical-based SMOS DisPATChD −0.177 0.387 0.327 −0.062 0.387 0.344 −0.367 0.508 0.281 SMAP VTCI −0.128 0.317 0.265 − − − −0.140 0.402 0.244 SMOS VTCI −0.271 0.377 0.261 − − − −0.447 0.524 0.265 Radiometer-based SMAP SFIM 0.039 0.373 0.238 −0.007 0.370 0.202 −0.238 0.344 0.231 Oversampling-based SMAP EnhancedA −0.082 0.274 0.222 −0.074 0.225 0.213 −0.222 0.354 0.242 SMAP EnhancedD −0.103 0.290 0.224 −0.110 0.286 0.189 −0.222 0.354 0.242 SMAP PassiveA −0.086 0.312 0.267 −0.091 0.261 0.245 −0.259 0.362 0.262 Uniform field SMAP PassiveD −0.114 0.313 0.250 −0.158 0.297 0.233 −0.236 0.369 0.262 SMOS PassiveA 0.020 0.309 0.233 0.158 0.307 0.230 −0.178 0.368 0.248 SMOS PassiveD −0.124 0.389 0.230 −0.089 0.270 0.230 −0.407 0.497 0.248
Funders | Funder number |
---|---|
ESA SEOM | 4000118762/16/I-NB |
Ecological Society of America | |
Australian Research Council | LE0453434, DP140100572 |
Monash University Malaysia |
Keywords
- Disaggregation
- Downscaling
- High resolution
- Inter-comparison
- SMAP
- SMAPEx
- SMOS
- Soil moisture