Created by: Vienna University of Technology

Funded by:

Funded by:

...unfortunately not.

The **state of a geophysical parameter** can be retrieved by converting information that is contained...

- in the observed radiation emitted by an object on the Earth’s surface (passive case) or
- in observed reflected radiation (active case)

Geophysical Parameter: e.g. soil moisture

Observable Quantity: radiance measured by the antenna

The TU Wien method represents an inverse model.

Models what we would observe (measure) under certain geophysical conditions

Processes involved in the **propagation of the radiation** from the Earth’s surface to the antenna:

interactions of the waves with illuminated matter such as **soil, vegetation canopy** or **atmospheric particles**

$$Y = f(x, \Omega)$$

*
$Y$...observable quantities (i.e., radiance measured by the antenna)
$x$...geophysical parameters of interest (e.g., temperature or soil moisture)
$\Omega$...set of controllable measurement conditions (e.g., the wavelengths, viewing direction, time, sun position, polarization, ...)
$f(.)$...complex function that relates *

Inversion of the forward model ⇒ retrieval of the parameters of interest from the observables.

We observe certain parameter values (e.g. backscatter) - what are the geophysical conditions that we are actually interested in (e.g. soil moisture)?

$$\hat{x} = g(\hat{Y}, \hat{\Omega})$$

*
$\hat{x}$ ...estimates of the geophysical parameters obtained by g(.)
$g(.)$ ...complex function
$\hat{Y}$ ...actual measurements
$\hat{\Omega}$ ...approximations of the measurement conditions
*

developed at the Vienna University of Technology

physically motivated empirical change detection method

no iterative adjustment process needed - direct retrieval from backscatter measurements possible

model parameters calibration requires a multi-year radar backscatter archive ⇒ land cover, surface roughness...

Variations of the backscatter coefficient $\sigma^0$ are related to...

- surface roughness
- changes in vegetation
- variations in the soil water content

Datatsets: European C-Band Scatterometers ESCAT and ASCAT (multi-incidence angle and multi-beam viewing capability)

- Incidence angle $\theta$: strong impact on the backscatter coefficient $\sigma^0$
- characteristic for roughness and vegetation
- soil moisture changes are not/only minimally affected

- Coarse resolution of the scatterometer (~ tenths of kilometres, see ASCAT Dataviewer screenshot showing
Austria and surrounding area)

⇒ roughness and land cover can be assumed temporally invariant

- $\sigma^0$ decreases/increases with vegetation change (more vegetation ⇒ higher $\sigma^0$)

⇒ the $\sigma^0$ time series changes in accordance to the vegetation phenology over time - Vegetation phenology cycle: seasonal scale

⇒ local short-term variability is negligible due to low resolution of the sensor - The relationship between soil moisture and $\sigma^0$ - expressed in [dB] - is linear

- Incidence angle at which $\sigma^0$ is rather stable in face of vegetation changes: "crossover angles"
- $\theta_{dry}$: ~ 25°
- $\theta_{wet}$: ~ 40°

- Normalisation of ESCAT backscatter measurements
- Azimuth angle normalisation
- Incidence angle normalisation

Vegetation correction

Soil moisture estimation

Soil moisture uncertainty estimation

...each of these steps is described in a respective lecture!