Soil moisture (and vegetation?) remote sensing products in Oklahoma

Size: px

Start display at page:

Download "Soil moisture (and vegetation?) remote sensing products in Oklahoma"

Horace Wheeler
5 years ago
Views:

1 Soil moisture (and vegetation?) remote sensing products in Oklahoma Jason Patton Plant and Soil Sciences, Oklahoma State University Wednesday, November 12, 2014 Oklahoma Workshop on Remote Sensing Technology and Applications

2 The coupling between weather/climate and soil moisture is apparent in models. (Koster et al. 2004)

3 Current regular soil moisture measurements are made at single points. 30 cm CS650 ~10 cm (Campbell Scientific)

4 Point measurements may not represent larger scale averages of soil moisture m volumetric soil moisture m (Bramer et al. 2013)

5 Weather and climate models need soil moisture data for initialization and validation at large spatial scales (>1 km), while in-situ measurements are available at point (~10 cm) scales. Satellite remote sensing of soil moisture can provide global measurements of soil moisture at large spatial scales.

6 Outline I. Soil Moisture and Ocean Salinity mission II. Soil Moisture Active Passive mission III. Cosmic-ray Soil Moisture Observing System

7 SMOS is the Soil Moisture Ocean Salinity satellite mission. European Space Agency Launched November 2009 Passive L-band (1.4 GHz, 21 cm) 43 km average resolution (ESA) Sensitive to top 3-5 cm of soil Polar orbiting: Measurements every 3 days at equator More often at higher latitudes

8 The tau-omega model describes the natural emission of microwave radiation from Earth s surface. T B = T soil (1 R soil ) e /µ (1) +(1 e /µ )(1!)T veg (2) +(1 e /µ )(1!)T veg R soil e /µ (3)

9 The tau-omega model describes the natural emission of microwave radiation from Earth s surface. T B = T soil (1 R soil ) e /µ (1) +(1 e /µ )(1!)T veg (2) +(1 e /µ )(1!)T veg R soil e /µ (3) (1) (2) (3)

10 The tau-omega model describes the natural emission of microwave radiation from Earth s surface. T B = T soil (1 R soil ) e /µ (1) +(1 e /µ )(1!)T veg (2) +(1 e /µ )(1!)T veg R soil e /µ (3) (1) (2) (3) Rsoil is soil reflectivity Rsoil = f(soil moisture, roughness)

11 The tau-omega model describes the natural emission of microwave radiation from Earth s surface. T B = T soil (1 R soil ) e /µ (1) +(1 e /µ )(1!)T veg (2) +(1 e /µ )(1!)T veg R soil e /µ (3) (1) (2) (3) Rsoil is soil reflectivity Rsoil = f(soil moisture, roughness) τ is vegetation optical thickness τ = f(vegetation water content) = b veg water content (VWC)

12 The tau-omega model describes the natural emission of microwave radiation from Earth s surface. T B = T soil (1 R soil ) e /µ (1) +(1 e /µ )(1!)T veg (2) +(1 e /µ )(1!)T veg R soil e /µ (3) (1) (2) (3) Rsoil is soil reflectivity Rsoil = f(soil moisture, roughness) τ is vegetation optical thickness τ = f(vegetation water content) = b veg water content (VWC) ω = Veg Scattering Albedo

13 SMOS uses a multi-angular approach to simultaneously estimate soil moisture and optical thickness. SMOS

14 SMOS uses a multi-angular approach to simultaneously estimate soil moisture and optical thickness. SMOS

15 SMOS uses a multi-angular approach to simultaneously estimate soil moisture and optical thickness. SMOS

16 SMOS uses a multi-angular approach to simultaneously estimate soil moisture and optical thickness. SMOS also assumes ω = 0.

17 Validation of SMOS has, so far, shown a slight dry bias in most cases, but it does capture dynamics well. See: Al Bitar et al. 2012; Gherboudj et al. 2012; Collow et al. 2012; Magagi et al. 2013

19 Vegetation optical thickness from SMOS is very noisy, but still may contain some information about vegetation FDA smoothed 21-day avg Apr May Jun Jul Aug Sep Oct Nov Dec day of year (Patton, 2014)

20 Vegetation optical thickness from SMOS is very noisy, but still may contain some information about vegetation FDA smoothed 21-day avg Δτ Apr May Jun Jul Aug Sep Oct Nov Dec day of year (Patton, 2014)

21 The change in τ over the growing season can be related to county crop yield estimates (in Iowa). county crop yield [kg m 2 ] m(2010) = county

22 The change in τ over the growing season can be related to county crop yield estimates (in Iowa). county crop yield [kg m 2 ] county crop yield [kg m 2 ] m(2010) = county m(2011) = county

23 The change in τ over the growing season can be related to county crop yield estimates (in Iowa). county crop yield [kg m 2 ] county crop yield [kg m 2 ] m(2010) = county 0.3 county crop yield [kg m 2 ] m(2011) = county m(2012) = county (Patton, 2014)

24 When using SMOS data, be considerate of the sensing depth, noise, radio frequency interference, and the grid spacing. South Fork SMOS soil moisture m 3 m 3 residuals m 3 m rate of soil drying m 3 m 3 day South Fork SMOS days after rain (Rondinelli et al, in review)

25 When using SMOS data, be considerate of the sensing depth, noise, radio frequency interference, and the grid spacing cm fit 4 6 cm fit 0.32 soil moisture, m 3 m time since last rainfall, days (Rondinelli et al, in review)

26 When using SMOS data, be considerate of the sensing depth, noise, radio frequency interference, and the grid spacing.

27 When using SMOS data, be considerate of the sensing depth, noise, radio frequency interference, and the grid spacing. SMOS pixels in Oklahoma x 15 km x km

28 When using SMOS data, be considerate of the sensing depth, noise, radio frequency interference, and the grid spacing. SMOS pixels in Oklahoma 43 km km

29 When using SMOS data, be considerate of the sensing depth, noise, radio frequency interference, and the grid spacing. SMOS pixels in Oklahoma 43 km km

30 SMOS data L1 (brightness temps) and L2 (soil moisture, tau) available from ESA: Proprietary format, use BEAM or Matlab API (req. 64-bit Linux) to view or convert to more friendly formats: BEAM: Matlab Read API: L3 (3-day/monthly soil moisture) available from CATDS in NetCDF: Some L2 (soil moisture & tau only) available from the Iowa Environmental Mesonet:

31 SMAP is the Soil Moisture Active Passive satellite mission. NASA Launching January 2015 Active and Passive L-band 3 km and 36 km resolutions 10 km combined active/passive Sensitive to top 3-5 cm of soil (NASA) Polar orbiting: Measurements every 3 days at equator More often at higher latitudes

32 The tau-omega model describes the natural emission of microwave radiation from Earth s surface. T B = T soil (1 R soil ) e /µ (1) +(1 e /µ )(1!)T veg (2) +(1 e /µ )(1!)T veg R soil e /µ (3) (1) (2) (3) Rsoil is soil reflectivity Rsoil = f(soil moisture, roughness) τ is vegetation optical thickness τ = f(vegetation water content) = b veg water content (VWC) ω = Veg Scattering Albedo

33 The main difference between SMOS and SMAP passive soil moisture retrieval is multi-angle vs. single angle approach. SMAP

34 The main difference between SMOS and SMAP passive soil moisture retrieval is multi-angle vs. single angle approach. SMAP

35 The main difference between SMOS and SMAP passive soil moisture retrieval is multi-angle vs. single angle approach. SMAP

36 SMOS SMAP L-band L-band Passive-only (43 km pixels) Multi angle, retrieves τ RFI plagued in regions 15 km ISEA grid (oversampled) Already in orbit Radar dissaggregation of passive pixels (36 km to 10 km) Single angle, requires τ RFI mitigation built in EASE-Grid 2.0 (grid spacing approx. matches resolution) Yet to be launched

37 The baseline SMAP soil moisture retrieval algorithm will require an outside source of vegetation data, will use an NDVI climatology to estimate τ. NDVI > Vegetation Water Content (VWC) > τ VWC = NDVI NDVI + stem factor NDVI max NDVI min 1 NDVI min = b VWC, (SMAP L2 Passive ATBD)

38 Under this baseline approach, SMAP may not be sensitive to interannual variability in vegetation SMAP Climo SMOS 2010 SMOS 2011 SMOS 2012 SMOS Mar Apr May Jun Jul Aug Sep Oct Nov Dec day of year (Patton, 2014)

39 SMAP data All products will be available through NSIDC in HDF-5 format about a year after launch. (NASA)

40 COSMOS is the Cosmic-ray Soil Moisture Observing System NSF project based out of University of Arizona Passive neutron counting sensor 700 m sensing area Sensitive to top cm of soil (dependent on soil moisture) (COSMOS)

41 COSMOS is being deployed across the US, already two (project-sanctioned) sensors in Oklahoma.

42 COSMOS counts fast neutrons, which are related to soil moisture because hydrogen slows neutrons. n v p

43 COSMOS counts fast neutrons, which are related to soil moisture because hydrogen slows neutrons. v n p v

44 COSMOS counts fast neutrons, which are related to soil moisture because hydrogen slows neutrons. n v p n n n p p n

45 COSMOS counts fast neutrons, which are related to soil moisture because hydrogen slows neutrons. v n n p n n p p v n

46 COSMOS requires careful calibration, which can change over time in areas with large changes in vegetation water content. (Hornbuckle et al 2012)

47 COSMOS data, publications, etc.

48 Thank you

Comparison of SMOS vegetation optical thickness data with the proposed SMAP algorithm

Graduate Theses and Dissertations Iowa State University Capstones, Theses and Dissertations 2014 Comparison of SMOS vegetation optical thickness data with the proposed SMAP algorithm Jason Carl Patton