update paper

docs/JOSS/paper.bib

@@ -181,4 +181,16 @@ @Article{FigaSaldana2002
   publisher = {Informa UK Limited},
 }
 
+@Article{Brocca2024,
+  author    = {Brocca, Luca and Barbetta, Silvia and Camici, Stefania and Ciabatta, Luca and Dari, Jacopo and Filippucci, Paolo and Massari, Christian and Modanesi, Sara and Tarpanelli, Angelica and Bonaccorsi, Bianca and Mosaffa, Hamidreza and Wagner, Wolfgang and Vreugdenhil, Mariette and Quast, Raphael and Alfieri, Lorenzo and Gabellani, Simone and Avanzi, Francesco and Rains, Dominik and Miralles, Diego G. and Mantovani, Simone and Briese, Christian and Domeneghetti, Alessio and Jacob, Alexander and Castelli, Mariapina and Camps-Valls, Gustau and Volden, Espen and Fernandez, Diego},
+  journal   = {Frontiers in Science},
+  title     = {A Digital Twin of the terrestrial water cycle: a glimpse into the future through high-resolution Earth observations},
+  year      = {2024},
+  issn      = {2813-6330},
+  month     = mar,
+  volume    = {1},
+  doi       = {10.3389/fsci.2023.1190191},
+  publisher = {Frontiers Media SA},
+}
+
 @Comment{jabref-meta: databaseType:bibtex;}

docs/JOSS/paper.md

@@ -27,9 +27,8 @@ bibliography: paper.bib
 
 The `rt1_model` package implements a generic solution to the radiative transfer equation applied to the problem of a rough surface covered by a tenuous distribution of particulate media as described in @Quast2016.
 
-It provides a flexible, object-oriented interface to specify the scattering characteristics of the ground surface and the covering layer via parametric distribution functions and to evaluate the resulting backscattering-coefficient ($\sigma_0$) for monostatic or bistatic measurement geometries as illustrated in Figure \ref{fig_model}.
-
-The underlying calculations are implemented via symbolic expressions, allowing the user to create fully customized model parameterizations. To speed up parameter retrieval stragegies, analytic solutions for the Jacobian of arbitrary model parameters are provided.
+It provides a flexible, object-oriented interface to specify the scattering characteristics of the ground surface and the covering layer via parametric distribution functions.
+The resulting model can then be evaluated to obtain backscattering-coefficient ($\sigma_0$) contributions for monostatic or bistatic measurement geometries up to first order as illustrated in Figure \ref{fig_model}. To speed up parameter retrieval stragegies, analytic estimates for the (zero-oder) Jacobian with respect to arbitrary model parameters are also provided.
 
 The package utilizes a minimal set of core dependencies, namely: `numpy` @Harris2020, `scipy` @Virtanen2020, `sympy` @Meurer2017 (with optional `symengine` @symengine support) and a set of visualizations created with `matplotlib` @Hunter2007.
 
@@ -38,28 +37,18 @@ The package utilizes a minimal set of core dependencies, namely: `numpy` @Harris
 # Statement of need
 
 
-Radiative transfer theory is used in a variety of contexts to retrieve biophysical characteristics from radar signals. The `rt1_model` package was developed to study soil-moisture retrievals from incidence-angle dependent backscatter measurements in the microwave-domain, provided for example by the ASCAT scatterometer onboard the METOP satellites @FigaSaldana2002 or synthetic aperture radar (SAR) instruments such as Sentinel-1 @Torres2012, or the upcoming NISAR @Rosen2021 mission.
-
-To date, the RT1 modeling framework was used for soil-moisture retrieval from microwave c-band radar data @Quast2019, @Quast2023 and adapted for rice-crop monitoring with a ground based bistatic scatterometer @Yadav2022.
-
-The bi-static nature of the distribution functions used in the parametrization of soil- and vegetation characteristics furthermore enables potential application to bi-static measurements as provided by GNSS systems or constellations such as SAOCOM @Scipal2017 or the upcoming Sentinel-1 companion mission Harmony @ESA2022.
-
-
-## Distribution functions
-
-The package provides a set of distribution functions (Isotropic, Rayleigh, HenyeyGreenstein, ...) that can be used to describe basic volume- or surface scatternig behaviors. More complex scattering scenarios can then be modelled by utilizing parametric linear-combinations.
+Radiative transfer theory is used in a variety of contexts to retrieve biophysical characteristics from radar signals. The `rt1_model` package was developed to study soil-moisture retrievals from incidence-angle dependent backscatter measurements in the microwave-domain as provided by the ASCAT scatterometer onboard the METOP satellites @FigaSaldana2002 or synthetic aperture radar (SAR) instruments such as Sentinel-1 @Torres2012, or the upcoming NISAR @Rosen2021 mission. It represents a generalization of the widely used "water cloud model" @Attema1978 to provide an intuitive bi-static parameterization framework that can be evaluated up to first-order (e.g. double-bounce) contributions.
 
-To support possibly anisotropic scattering characteristics, all functions are furthermore implemented with respect to a generalized scattering angle @Lafortune1997:
+The RT1 modeling framework was used to perform soil-moisture retrievals from microwave c-band radar data @Quast2019, @Quast2023, @Brocca2024 and adapted to study rice-crop monitoring with a ground based bistatic scatterometer @Yadav2022.
 
-$$\cos(\Theta_a) = a_0 \cos(\theta) \cos(\theta_s) + \sin(\theta)\sin(\theta_s) [ a_1 \cos(\phi)\cos(\phi_s) + a_2 \sin(\phi) \sin(\phi_s)]$$
+The bi-static nature of the model parameterization strategy furthermore facilitates potential application to bi-static observations provided by Global Navigation Satellite Systems (GNSS) or satellite constellations such as SAOCOM @Scipal2017 or the upcoming Sentinel-1 companion mission Harmony @ESA2022.
 
-where ($\theta, \phi$) denote the incident azimuth and polar angle and $(\theta_s, \phi_s)$ the corresponding exit angles and $(a_0, a_1, a_2)$ are the generalization parameters.
+## Parameterization strategy
 
-For example, a surface scattering distribution function that consists of a peak in specular direction and a second peak in incidence direction can be defined as:
+To allow modelling of a wide variety of scattering characteristics for the ground surface and the covering layer, the package provides a set of distribution functions (Isotropic, Rayleigh, HenyeyGreenstein, ...). All functions are hereby implemented with respect to a generalized scattering angle @Lafortune1997 to support modelling of anisotropic effects.
+In addition, parametric linear-combinations can be used to model more complex scenarios.
 
-$$BRDF = w * HG(-t, a_0=-1) + (1-w) * HG(t, a_0=1) \quad \textrm{with} \quad w, t \in (0,1)$$
-
-and implemented via:
+For example, to model a surface scattering behavior that consists of a peak in specular direction and a second peak in incidence direction, we can use a linear-combination of two HenyeyGreenstein functions:
 
 ```
 from rt1_model import surface
@@ -69,16 +58,19 @@ SRF_2 = surface.HenyeyGreenstein(t="t",  a=[ 1, 1, 1], ncoefs=12)
 SRF = surface.LinComb([("w", SRF_1), ("1-w", SRF_2)])
 ```
 
-
-![Linear Combination of surface BRDFs. \label{fig_lin_comb_brdf}](lin_comb_brdf.png)
+![Linear Combination of surface BRDFs. \label{fig_lin_comb_brdf}](lin_comb_brdf_latex.png)
 
 ## Parameter Retrieval
 
-The documentation provides a set of retrieval examples...
+To show potential applications of the `rt1_model` package to perform forward-simulations as well as parameter retrievals from monostatic measurements, the documentation provides several examples on how to use the package in conjunction with non-linear least squares regression optimization provided by `scipy.optimize`.
 
-![Example of the analyzer-widget for a RT1 model result. \label{fig_retrieval_static}](retrieval_static.png)
+Examples hereby show how to perform "static" parameter retrievals (Figure \ref{fig_retrieval_static}) as well as multi-temporal timeseries optimization (Figure \ref{fig_retrieval_multi_temporal})).
 
+The provided analyzer-widget can then be used to assess effects of model parameters on the individual backscattering contributions.
+
+![Example of the analyzer-widget for a RT1 model result. \label{fig_retrieval_static}](retrieval_static.png)
 
 
+![Example of optimization results for a multi-temporal retrieval. \label{fig_retrieval_multi_temporal}](multi_temporal_retrieval.png)
 
 # References