| 28 Jul 2026
Representing prescribed fire and mechanical thinning in a demographic vegetation model: mechanisms driving forest structure, fuel, and demographic rates in California mixed-conifer forests
Abstract. Prescribed fire and mechanical thinning are common fuel management and forest restoration practices in fire-prone ecosystems, particularly where long-term fire suppression has led to significant fuel accumulation and altered stand structure. Existing field evidence shows varied outcomes depending on environmental and stand conditions, with long-term, landscape-scale management effects remaining poorly understood. The majority of large-scale models that project long-term effects lack mechanistic management processes and are rarely validated for the ecological processes driving outcomes. To address this, we implemented a new prescribed fire model and adapted a pre-existing wood harvest model to represent restoration thinning in the Functionally Assembled Terrestrial Ecosystem Simulator (FATES), a dynamic vegetation demography model. We developed a model validation framework emphasizing the key ecological mediators and expected feedback between management, vegetation, and fire. Running simulations with varying weather conditions under which prescribed fire can occur (burn window) and with the forest thinning model across three sites in California’s mixed conifer forests, we found that FATES successfully captured the observed relative changes in stand structure, tree size distribution, fuel load, and demographic rates following forest health treatments. Additionally, model results aligned with general ecological expectations and reproduced the key differences between thinning and prescribed fire: logging immediately and significantly reduces stem density and basal area but not surface fuel load, while prescribed fire results in nonsignificant to moderate reduction in stem density and basal area but an immediate decrease in surface fuel load. Notably, simulations with a wide, inclusive burn window resulted in cooler fires and caused less change in stem density and basal area than those with a narrow, exclusive burn window, but only at the warmest and driest modeled study site. Our work advances the integration of management activities into Earth system models by introducing a mechanistic, interactive prescribed fire model, and a validation framework grounded in empirical data to assist model evaluations and applications. The model provides a promising tool for projecting the complex climate-vegetation-fire-human interactions in dry conifer forests of the Western U.S. and other fire-adapted ecosystems with similar management history.
General comments:
This paper presents a new development and evaluation of two vegetation management modules in a vegetation demography model. Specifically, they present a Prescribed Burn module, and readapt a thinning module, of the FATES model, a model that can be further used in Earth system models (and thus has a great potential for exploring fire-climate-vegetation feedbacks). They apply the model using actual data from three California Sierra Nevada mixed-conifer sites to evaluate its performance. They present and discuss results according to the three sites, the management objectives, and different burning windows options. The strongest contribution is the ability to use a prescribed burning process that emerges from weather, management prescriptions and vegetation conditions.
The manuscript is clearly written, easy to follow, and addresses an important process usually overlooked by vegetation models (prescribed burn simulation and effects).
There are two main concerns I have about the paper are how the model is presented and evaluated:
are introduced throughout Sections 2.2-2.5, but there is no unified framework describing how these parameters were estimated, or calibrated. A dedicated parameterization section, including a table, methodology (sensitivity analysis, optimization, other sources, etc.) and the final values for the present paper would greatly improve reproducibility and facilitate application of the model.
Specific comments
L228–239: This section is difficult to follow. Is there a spatial conceptual figure (in the supplementary material) that could help clarify the fractional part of the prescribed burn module?
In the discussion, could the authors elaborate a bit more on how the model could be applied to other areas? Which parameters would need to change?
Mortality is poorly captured, and although the authors provide reasoning for this in the discussion, please revise the language used elsewhere in the manuscript where this process is described as being correctly captured.
In some cases, the authors attribute differences to differences in initial conditions. The authors should find a way to address this in order to achieve a more robust evaluation of the model.
L466–469: Why is there less thinning in the control than in the treated units in the observations? And why does the model also reproduce this pattern?
Figure 7: Please standardize the terminology used throughout the manuscript — litter / dead leaf / dead litter.
Figure 8: Why is sugar pine not included in the plot? Please explain.
L520–522: Please be more descriptive here — the current language is too soft. For instance, in some cases the model does not perform well (mortality).
L563–567: Fire practitioners rarely intend to affect the trees, but it is encouraging that the model is able to capture some mortality, especially if it is sensitive to weather conditions. However, this also explains why mortality is overpredicted.
Supplemental Figure S11: The y-axis differs considerably between burning windows. Is this correct — that fir mortality is roughly 50 times higher in the hot burn than in the cool burn? Please explain what is driving this difference.