Is LPJ-GUESS a version of LPJ?

No. These two models share certain features and were developed in parallel originally, but exhibit a number of fundamental differences, particularly regarding the representation of vegetation dynamics and (from LPJ-GUESS release version 3.0) nitrogen cycling. Intercomparison studies reveal that there can be substantial differences between the models in terms of performance and responses to drivers.

Is there a publication that fully describes the whole model?

No single publication includes a full description of every concept/equation/parameter. A fairly comprehensive description of the LPJ-GUESS 3.0 release version is available in Smith et al. (2014).

Does LPJ-GUESS include nitrogen?

Yes, from LPJ-GUESS release version 3.0. See Smith et al. (2014). Emissions of N-based trace gases is still under development.

What about phosphorus?

No, maybe in a future version.

Crops, managed forests and land use?

Yes, from LPJ-GUESS release version 4.0. See Lindeskog et al. (2013).

Methane, Arctic PFTs, permafrost?

Yes, in a customised “northern peatlands” version still partially under development. Some results may be seen in this paper and this one.

BVOCs?

Isoprene and monoterpene emissions from vegetation are included from release version 3.0. See Schurgers et al. (2009) and Arneth et al. (2011).

Does LPJ-GUESS use fixed PFT parameters or traits?

It does not resample data from a trait database like some recent models, but, as in other DGVMs, some key PFT traits, such as rubisco capacity, root:shoot ratio and biomass turnover, are simulated prognostically based on external drivers and realised ecosystem state, e.g. in accordance with ecological optimality theory.

LPJ-GUESS is a gap model, right? Can it really simulate non-forest vegetation?

Like other second-generation DGVMs, LPJ-GUESS simulates age-structured dynamics (demography) of woody vegetation as the emergent outcome of growth and competition for light, space and soil resources among individuals and a herbaceous understorey within replicate patches representing “random samples” of a simulated landscape. This approach is borrowed from so-called gap models. Unlike traditional gap models, LPJ-GUESS employs process-based representations of canopy gas exchange, plant resource use, growth and phenology, soil biogeochemistry and ecosystem hydrology. Unlike gap models, which are usually applied to forest vegetation at local scale or across a limited region, LPJ-GUESS uses generic parameterisations that allow it to be applied without recalibration to any climate zone or biome, including non-forest vegetation such as Arctic tundra or savannah.

Can it do individual species?

Yes, for example European and North American trees.

What's the time step?

Normally a day or optionally a month for fast processes, and a year for slow processes. From release version 3.0, LPJ-GUESS can also be configured to simulate a diurnal cycle for processes relating to canopy gas exchange.

Is LPJ-GUESS suitable for coupling to an atmospheric model?

LPJ-GUESS has been coupled to a regional climate model to account for vegetation biophysical feedbacks, see Smith et al. (2011). It is currently being set up to account for vegetation and carbon cycle dynamics in a global Earth system model.

Is there source code available for download?

Source code is normally made available on request to bona fide research users. Conditions apply in the case of model versions still under active development. See contacts.

Is there a simple demo version I can try out?

LPJ-GUESS Education (Windows executable, no source code) provides an easy way to get an impression of the capabilities and behaviour of the model. It is not suitable for research use.

Does the model come with all required input/driver data?

No, normally you need to provide your own climate, soils and CO2 input data, and implement your own code for reading the data. By agreement we can sometimes help provide driver data sets and input/output code for historical (hindcast) simulations.

Yikes, you mean I have to do some of the programming myself?

Most probably, yes. The software comes with documentation that should help and the source code itself is fairly well structured and commented. However, some basic skills in C/C++ are a minimum requirement.

What language is the source code written in?

ANSI C++

Will it compile/run on Windows? Linux? Mac?

The software is designed to be portable and has been compiled and run on a variety of Linux/Unix systems as well as Windows PCs and Macs.

What compiler do I need?

Any modern C++ compiler should do. The software comes with scripts invoking cmake to automate compilation and building from source. Cmake should detect the C++ compiler and other relevant aspects of the environment of your system.

How long will a simulation take to run?

A global run at half-degree resolution (about 50,000 grid cells) with an acceptable minimum of 5 replicate patches might take around 24 hours on a modern (2014) state-of-the-art PC. Divide by [5×50,000] to get simulation time per grid cell and patch. If you are planning to do global runs you may need to consider going to an HPC platform (supercomputer cluster).

Can I run LPJ-GUESS in parallel mode on my HPC system?

We routinely run the model in parallel on a number of different Linux cluster systems. The software comes with scripts to compile and run the model as a parallel application on systems implementing MPI and OpenPBS.