HannesDeDeurwaerder/iSWIFT: iSWIFT: Inverse method to derive the relative distribution of the absorbing root surface using the SWIFT model

This package provide an inverse approach to estimate the distribution of the absorbing root surface from isotope data and plant hydraulic traits considered. While this methodology can be applied on all models combining a mixing model approach with plant hydraulic models, here it is applied on the SWIFT model (De Deurwaerder et al. 2020, Biogeosciences). 'iSWIFT' provide a set of new function, with at its core the 'LogLikOptim' function which optimizes the log-likelihood function in order to obtain the best estimate of beta (beta.hat in De Deurwaerder et al (submitted)). Random set of SWIFT variables are generated in the 'VarMatrix' module where values are randomly extracted from user defined ranges. 'SoilHeterogeneity' will generate synthetic varibility in the soil water potential and soil water isotope composition gradients with depth for soil profile fitting, also required by the SWIFT model. SWIFT model outputs, with extraction error, are subsequently generated in the 'ProbabilitySpace' function, providing synthetic field data for which conditional desity probability distribution are calculated via the 'CDPD'-function, required to evaluate the log likelihood of field data with the SWIFT model generated data for a specific beta within the 'LogLik' function. Note that since the iSWIFT package builds further on the SWIFT package, it contains the same functionalities and requires the same variable and parameter declarations as decribe in more detail in the 'SWIFT' package.The SWIFT model functionalities preserved are: Function 'SWIFT_SB' and 'SWIFT_H' can respectively be used to calculate the isotopic signature over time at the stem base of the plant, or at a by the user defined height and time. Additional functions are 'SoilRootCond', 'PSI0calc' and 'Bprep' which in this order help the user to define (i) the soil to root conductivity for every defined soil layer, (ii) the water potential at stem base for every timestep, and (iii) the root length distribution for every defined soil layer.

install.packages("remotes")
remotes::install_github("HannesDeDeurwaerder/iSWIFT")