This repository provides the Vaganov-Shashkin tree-ring growth model (VSM) in MATLAB (Octave-compatible). VSM, originally written in Fortran, mimics subdaily and daily resolution processes of cambial growth in trees as a function of soil moisture, air temperature, and insolation, with environmental forcing modeled as the principle of limiting factors. Our re-implementation in a high level interpreted language, while sacrificing speed, provides opportunities to systematically evaluate model parameters, generate large ensembles of simulated tree-ring chronologies, and embed proxy system modeling within data assimilation approaches to climate reconstruction. We also provide examples of model applications which permit process-level understanding of tree ring width variations in response to environmental variations and boundary conditions.
Please cite this work when using this model:
Anchukaitis, K.J., M.N. Evans, M. K. Hughes, and E. Vaganov, An interpreted language implementation of the Vaganov-Shashkin tree-ring proxy system model, Dendrochronologia, 60, 125677, doi:10.1016/j.dendro.2020.125677, 2020
A preprint is available at EartharXiv.
Please inform Kevin Anchukaitis if you identify any bugs.
The code model function is vsm. This function runs both the Environmental and Growth blocks using daily temperature and precipitation input as well as latitude (for daylength calculations). The function call is:
[output] = vsm(T,P,phi,syear,eyear,parameters,varargin)where T is the daily temperature (in ), P is the daily precipitation (in ), phi is the latitude of the tree-ring site or the latitude of the source of the daily meteorological data, syear is the starting year of the simulation (the first year of the daily meteorological data), eyear is the last year of the simulation, and parameters is the name of the structure containing the parameters. varargin allows addition variables to be passed to the model without affecting the standard, required set (the first 6 input variables). output is a structure containing the annual, daily, and cell-specific outputs from the simulation.
vsm expects an input structure parameters that contains the tunable parameters that determine model behavior. Several example parameter sets are included in this distribution (generic_parameters,wm1_parameters,e06_parameters, and a06_parameters)
Although the code has been built to attempt to catch errors related to missing input data, you should estimate or fill missing temperature and precipitation data values prior to running vsm. You may use built-in MATLAB functions like fillmissing, but we also include a utility function vsm_fillmiss that mimics the internal missing value filling in the original FORTRAN code.
Initial testing of the code in this repository was primarily done on a Macbook Pro (Retina, 15-inch, Mid 2015, 2.8 GHz Intel Core i7, 16 GB 1600 MHz DDR3) running MATLAB R2017b on MacOS 10.13.6 (High Sierra). A single 22 year simulation (as for the White Mountain bristlecone example) takes approximately 0.20 seconds to complete. A ~80 year simulation (as for the Mohonk Hemlock simulation) take approximately 1.1 seconds to complete.
This repository comes with three demonstration files that recreate the figures in the manuscript and demonstrate the functionality of vsm. The first (dendrochronologia_demo_1.m) reproduces the White Mountain (USA) bristlecone simulations from the manuscript. This script contains a loop over a range of drainage rate parameter value that can take up 40 second or more to complete depending on your system, but the code is annotated if you wish to reduce the number of drainage rates used for the simulation (default is 191 different rates). The second script (dendrochronologia_demo_2.m) reproduces the analysis of the temperature parameter sensitivity at Mohonk (USA) using a Latin Hypercube design (lhsdesignbnd). The manuscript itself uses a design with 1000 draws, but this can require anywhere from 15 to 30 minutes to complete, do the file is set to start with a design of 100 members initially. The third demonstration script uses syntax for setting the characteristics of bar charts introduced in MATLAB R2014B.
The original FORTRAN model was developed for Northern Hemisphere applications. As a consequence the definition of a year in the model is from January 1st to December 31st, which is unlikely to be adequate for Southern Hemisphere tree-ring applications.
The core model function vsm works in Octave without modification (avoiding use of nansum and nanmean that are not part of core Octave distributions). Some of the demo scripts and the helper functions therein, however, use commands not available or with a different syntax in Octave and in general Octave (testing on version 4.4.1 running on a Macbook Pro with Darwin Kernel Version 17.7.0) was found to be substantially slower than the same operations in MATLAB.
This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/4.0/ or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.