Skip to main content
CTrees Logo
  • Land Carbon

Forest productivity

Estimate the aboveground biomass growth rate for any naturally regenerated forest globally

Lead scientist

This dataset is available by partnership with CTrees

Carbon growth rate

Introducing forest productivity

CTrees’ forest productivity dataset models annual carbon removal rates from forest restoration activities. It estimates the aboveground biomass (AGB) growth rate (MgC ha⁻¹ yr⁻¹) at a 100-meter resolution for any naturally regenerated forests at age zero to 100.

CTrees provides the productivity rate estimates globally at 100-meter resolution, allowing for calculation of annual carbon removal rates. The estimates can then be used in ARR and IFM carbon projects to predict potential carbon sequestration in a project’s lifetime.

While traditional methods rely on ground-based measurements, CTrees’ approach applies remote sensing to better represent the spatial and temporal variability of global forests, particularly young forests, which are often underrepresented in calculations. The CTrees model brings improved accuracy by training spatially explicit, locally trained growth curves for each 100-meter cell.

Annual data on forest carbon removal

Units

MgC ha-1 yr-1

Spatial Resolution

100 meters (m)

Temporal Resolution

Annual, reflecting static climatic and biophysical conditions

Temporal Domain

100 years of forest stand age, based on average conditions between 1980-2020

Data Type, Scale, Offset
  • Amax: UInt16; Scale: 0.1; Offset: none
  • k: UInt16; Scale: 0.0001; Offset: none
  • m: UInt16; Scale: 0.001; Offset: none
File Format

Geotiff, cloud optimized GeoTIFF

Methodology

Forest productivity (MgC ha-1 yr-1) is defined as the annual change of AGB, calculated via the “stock change” method. Forest AGB accumulation over time was modeled using the locally trained Chapman-Richards (CR) growth function (Eq 1):

AGBt = Amax * (1-b * exp(-t * k)) m

Where: Amax, k, and m are the parameters of CR growth function. AGBt represents the AGB density at a given forest age (t). Parameter Amax is the asymptote, indicating the maximum potential AGB density. Parameter b influences the initial carbon density, and was set as 1 assuming all living forest biomass is cleared after a stand-replacing disturbance, or forest biomass start with 0 in reforestation activity. Parameter m is a shape parameter, typically greater than 1, indicating an S-shaped growth curve with slower initial biomass recovery. Parameter k signifies the growth rate.

The curves extend to 100 years, reflecting static climatic and biophysical conditions, and are suitable for predicting carbon removals over the next 20-30 years. However, as climate conditions deviate outside historical conditions in the future, the productivity curves may not accurately predict carbon removals. The data is derived from long-term, multi-decade climate and biophysical data, as well as remote sensing layers that reflect average conditions between 1980-2020.

Input data

Input data includes (1) AGB and stand age data derived from a "space-for-time" approach integrating remote sensing observations (n = 16,677,616) and field plot data (n = 105,091); (2) a list of spatial predictors representing climatic, biophysical, and disturbance factors.

Modeling approach

Modeling forest AGB accumulation through time

Forest AGB accumulation through time was modeled using the locally trained CR growth function for each 1-degree grid cell with at least 10% of forest cover (n = 7,939), using AGB and stand age data derived from a “space-for-time” approach that integrates remote sensing observations (n = 16,677,616) and field plot data (n = 105,091).

The model fitting within each 1-degree grid cell resulted in three biologically meaningful growth function parameters (Amax, k, and m), which were then used as dependent variables during the next step.

The analysis differs from previous studies in two key ways. First, by integrating remote sensing observations with field data, we extend coverage into regions with limited ground-based measurements, enabling a more comprehensive representation of the spatial and temporal variability of global forests—particularly young forests, which are often underrepresented.

Second, unlike prior studies that applied global models to derive CR growth functions, we developed spatially explicit, locally trained growth curves for each 1-degree grid cell.

Downscaling CR growth function parameter to 100-meter resolution

Then, the Amax, k, and m for 1-degree grid cell was downscaled to 100-meter resolution using a machine learning approach. Specifically, a Random Forest (RF) regression model was fit for each parameter with climatic, soil, and biophysical predictors to map variability in AGB growth dynamics across each hectare of global regenerated forest.

Calculating AGB growth rates

The 100-meter resolution maps of CR growth function parameters enabled spatially explicit predictions of aboveground biomass carbon (AGC) storage through time for each hectare of global regenerated forest. Forest AGB growth rates were calculated using a “stock change” approach, which computes the annual change in AGC between consecutive years (MgC ha⁻¹ yr⁻¹, = AGCyear+1 - AGCyear). We calculated the AGB growth rates to be consistent with IPCC Guidelines for National Greenhouse Gas Inventories (NGHGI, e.g., Table 4.9 in (IPCC, 2019)).

Validation approach

Validation was conducted at various scales:

Regional-level AGB accumulation curves validation

We first compared modeled AGB accumulation curves with U.S. FIA data across 20 different forest types. To do so, we sampled the 100-meter CR growth function parameter maps at FIA plot locations (n = 29,883) and aggregated the data into 20 forest type groups. We then plotted both the mean and pixel-level model-derived AGB accumulation curves for each forest type, overlaid with FIA observations represented as boxplots at 10-year intervals. This approach allowed for a direct visual comparison between model-predicted and FIA-derived AGB trajectories.

Site-level AGB growth rate validation

Site-specific AGB growth rates were compared with published studies that inform the Intergovernmental Panel on Climate Change (IPCC) default growth rates and recent research. We prioritized sources reporting age-specific growth rates based on (1) site-level chronosequence measurements or (2) regional estimates derived from space-for-time substitutions or large forest plot networks. For site-specific studies, we extracted AGB growth rates at the reported geographic coordinates and matched them to corresponding forest ages. For regionally aggregated studies, we calculated the mean predicted growth rates within equivalent age ranges and geographic regions for direct comparison.

Data Access

Use cases

Applying the maps of Amax, k, and m, users can estimate forest biomass and AGB growth rates for stands of different ages at pixel, project, or regional levels. This estimation capability helps users address relevant questions for science and carbon market project design. For example, using the spatially explicit predictions of AGB storage over time for each hectare of global regenerated forest, one can determine the carbon sequestration potential of these forests by the year 2050. The model can also be used to quantify the carbon sequestration potential of an ARR project.

References

Liu, Z., Saatchi, S., Cook-Patton, S.C. et al., Recovering young forests dominate the current and future land carbon sink, In review.

Robinson, N., Drever, C.R., Gibbs, D.A. et al. Protect young secondary forests for optimum carbon removal. Nat. Clim. Chang. 15, 793–800 (2025). https://doi.org/10.1038/s41558-025-02355-5 (opens in a new tab)

Team

Page last updated Nov 16, 2026

All Datasets

Stay Connected

Subscribe to our monthly newsletter and be the first to know about new data releases, research, and events.

By submitting your email address, you agree to receive communications from CTrees. You can opt-out at any time.