Biomass estimation Drone Guide

# Biomass Estimation Drone Guide – Forestry

Introduction

Understanding how much woody material exists within a forest is important for commercial forestry, forest inventory, carbon management, conservation and long-term planning. Biomass information can support estimates of timber resources, forest growth, fuel availability and the amount of carbon stored within vegetation.

Traditionally, forestry biomass estimation relies heavily on field measurements. Sample plots are established and forestry professionals measure variables such as tree diameter, height, species and density. Established allometric models are then used to estimate biomass.

These methods remain essential because a drone does not directly weigh a forest.

What drones provide is the ability to collect detailed information about forest structure across much larger areas than individual ground plots.

LiDAR can measure canopy height and three-dimensional structure. Photogrammetry can generate canopy surface models, while multispectral imagery can provide additional information about vegetation characteristics. When these datasets are calibrated against reliable field measurements, statistical or machine-learning models can estimate biomass across a wider forest area.

This creates an important distinction:

Drones generally estimate forest biomass indirectly by measuring characteristics that are related to biomass and connecting those measurements with field data and validated models.

Used correctly, this can transform forest inventory from a limited collection of sample plots into a much more spatially detailed understanding of how biomass is distributed across the landscape.

From Tree Measurements to Forest Biomass

Forest biomass includes biological material contained within trees and other vegetation. Depending on the project, analysts may focus specifically on above-ground woody biomass or use a broader definition.

Above-ground biomass can include stems, branches, bark and foliage.

Below-ground root biomass may also be relevant to some carbon assessments, but it is significantly more difficult to measure directly and generally cannot be observed by ordinary aerial drone sensors.

Traditional forestry methods estimate above-ground biomass using relationships between measurable tree characteristics and biomass.

Tree diameter is particularly important.

Height, species and wood density may also contribute.

Field crews can measure these variables within representative plots and apply appropriate allometric equations.

The challenge is extending those plot-level measurements across an entire forest.

This is where remote sensing becomes valuable.

Drone data can describe forest height, canopy structure and spatial variation between field plots. Models can then use these relationships to estimate biomass across the wider survey area.

LiDAR and Three-Dimensional Forest Structure

LiDAR is one of the most powerful drone technologies for forestry biomass estimation because biomass is strongly related to forest structure.

A LiDAR sensor generates large numbers of three-dimensional measurements.

Returns may come from the upper canopy, branches, understory vegetation and ground.

Once ground points have been identified, the height of vegetation above the terrain can be calculated.

This allows the creation of canopy-height models.

LiDAR data can also describe vertical forest structure.

Metrics may be calculated for different height levels within the canopy, providing more information than simply measuring the tallest point.

In suitable forest conditions, individual tree crowns may also be segmented.

Tree-height estimates can then be generated.

These structural measurements can be compared with biomass values derived from field plots.

Statistical models or machine-learning methods can identify relationships between LiDAR metrics and measured biomass.

The resulting model can then estimate biomass across other parts of the forest covered by the LiDAR survey.

The accuracy depends heavily on forest type, field data, sensor quality, point density and modelling methodology.

Photogrammetry and Canopy Models

LiDAR is not the only technology available.

High-resolution RGB imagery can also support biomass-related forest measurements through photogrammetry.

Overlapping photographs are processed into a three-dimensional point cloud and Digital Surface Model.

If a reliable Digital Terrain Model is available, canopy height can be calculated by comparing the canopy surface with the underlying ground.

This can provide useful structural information.

Photogrammetry can be particularly effective in relatively open forests, plantations and environments where tree crowns are clearly visible.

Dense canopy creates limitations.

A camera cannot normally see the ground beneath closed vegetation.

If an accurate terrain model is not already available, calculating true tree height becomes more difficult.

One practical approach is to combine photogrammetry with an existing LiDAR-derived terrain model.

The terrain changes relatively slowly, while the canopy changes through growth and harvesting.

A detailed terrain model can therefore provide the geographic foundation for repeated lower-cost RGB surveys.

Individual Tree Detection and Measurement

Some forestry biomass workflows operate at the individual-tree level.

High-resolution imagery or LiDAR point clouds can be processed to identify tree crowns.

Algorithms may estimate tree location, height and crown dimensions.

This creates a digital inventory of individual trees.

In plantations with relatively regular spacing and consistent species, automated tree detection can work particularly well.

Complex mixed forests are more challenging.

Tree crowns overlap.

Understory trees may be hidden beneath larger trees.

Different species can have very different crown shapes.

Aerial systems may therefore underestimate the number of suppressed or lower-canopy trees.

Individual-tree analysis should consequently be validated against field measurements.

Where reliable segmentation is possible, however, it can provide extremely detailed information about the distribution of forest structure.

Multispectral Data and Vegetation Information

Multispectral sensors provide an additional information layer.

These cameras measure selected wavelength bands beyond ordinary visible imagery.

Vegetation indices can then be calculated.

Such indices can provide information related to canopy condition, vegetation density and photosynthetic activity.

When combined with structural information from LiDAR or photogrammetry, multispectral data may improve biomass models in some environments.

For example, two forest areas might have similar canopy height but different vegetation characteristics.

Spectral information may help the model distinguish them.

However, vegetation indices should not be treated as direct measurements of biomass.

Dense forests can also experience spectral saturation, where increasing vegetation no longer produces a proportionally large change in the index.

Multispectral information is therefore generally strongest when combined with structural and field data.

Field Plots and Calibration

Ground measurements remain one of the most important parts of drone biomass estimation.

Without reliable field data, it can be difficult to convert remote-sensing measurements into defensible biomass values.

Representative sample plots should capture the variation within the forest.

Field teams may record species, diameter at breast height, tree height and other relevant characteristics.

Appropriate allometric equations are then used to estimate biomass within each plot.

The corresponding drone measurements are extracted for those same areas.

A model can then be developed linking the aerial metrics with the field-derived biomass.

Part of the data should ideally be reserved for independent validation rather than using every plot to build the model.

This provides a more realistic indication of how the model performs on locations it has not already seen.

The quality of the field data can therefore be just as important as the quality of the drone sensor.

Forest Inventory and Timber Management

Biomass estimation can provide valuable information for commercial forest management.

Maps can show how forest structure varies across compartments.

Areas with higher or lower estimated biomass can be identified.

This can support inventory planning and help determine where additional ground measurements may be required.

Repeat surveys can also provide information about growth.

A plantation mapped at different stages may show increasing canopy height and structural development.

Harvested areas can be clearly identified.

Thinning operations can also produce measurable changes.

These datasets can support strategic planning.

However, biomass should not automatically be treated as merchantable timber volume.

Not all biological material is commercially usable timber.

Species, stem form, quality, diameter classes and market requirements must also be considered.

Traditional forestry inventory remains necessary when determining commercial timber resources.

Carbon Stock Estimation

Forest biomass is closely connected with carbon accounting because a substantial proportion of dry woody biomass consists of carbon.

Biomass estimates can therefore contribute to calculations of above-ground carbon stocks.

The process normally involves several stages.

Remote sensing estimates above-ground biomass.

Appropriate conversion factors or models are then used to estimate carbon.

Additional components may be required for a complete forest carbon assessment, including roots, deadwood, litter and soil carbon.

A drone cannot directly measure all of these carbon pools.

This distinction is important for carbon projects.

A high-resolution drone biomass map may provide excellent information about above-ground vegetation while still representing only part of total ecosystem carbon.

Carbon accounting should therefore follow the relevant methodology, validation requirements and professional standards applicable to the project.

Biomass Change and Forest Growth

One of the most valuable applications of drone monitoring is measuring change over time.

A single biomass estimate provides a snapshot.

Repeated surveys can show how forest structure develops.

LiDAR surveys taken several years apart can identify changes in canopy height and structure.

Photogrammetry may provide more frequent updates where a reliable terrain model already exists.

These datasets can support growth modelling.

They can also identify areas affected by harvesting, storm damage, wildfire or disease.

When combined with field measurements, repeated remote-sensing surveys can improve understanding of how biomass is changing across different parts of the forest.

This is particularly valuable because forest growth is not spatially uniform.

Soil, elevation, moisture, species and management all influence development.

A spatial biomass map reveals this variation much more clearly than a single estate-wide average.

Biomass After Storms, Wildfire and Disturbance

Major disturbances can rapidly alter forest biomass.

Storms may bring down large numbers of trees.

Wildfires can remove or damage vegetation.

Insects and disease can cause gradual canopy decline.

Drone surveys can document these changes.

RGB imagery can map visible disturbance.

LiDAR can quantify changes in canopy structure.

Multispectral imagery may provide supplementary information about vegetation condition.

Pre-event datasets are particularly valuable.

If a forest was mapped before a storm, the post-event survey can be compared directly with the earlier model.

This can help estimate the spatial extent of structural change.

However, fallen timber may still remain within the area as biomass even though the standing canopy has disappeared.

The definition of biomass being measured must therefore be clear.

Remote-sensing models designed for standing trees may not accurately represent large quantities of fallen woody material.

AI and Machine Learning

Biomass estimation is increasingly connected with machine learning.

Forest structure produces complex datasets containing many potential variables.

LiDAR may generate canopy height, percentile heights, point-density metrics and vertical-distribution information.

Multispectral imagery provides additional spectral variables.

Terrain, species and environmental information may also be available.

Machine-learning models can identify relationships between these variables and field-measured biomass.

This can improve estimation in some situations.

However, complex models are not automatically more accurate.

A model trained in one forest may perform poorly in another with different species, age classes or environmental conditions.

Independent validation is therefore essential.

AI should assist forestry professionals in extracting relationships from large datasets rather than creating an assumption that biomass can be determined automatically without reliable ground evidence.

Biomass Mapping in GIS

Once biomass estimates have been produced, they can be represented geographically.

GIS allows the forest to be divided into grids, stands or management compartments.

Each area can contain an estimated biomass value.

This produces a biomass-density map.

Forestry managers can then combine this information with species, age, terrain, roads, harvesting plans and environmental areas.

Historical biomass maps can also be stored.

The GIS becomes a record of how forest resources change through time.

Carbon information may form another layer.

For larger organisations, these datasets can become part of a forest digital twin containing inventory, terrain, infrastructure and environmental information.

Large-Area Surveys and Multi-Scale Monitoring

Drone surveys provide extremely high-resolution information, but very large forestry estates may be inefficient to monitor exclusively using small drones.

A multi-scale approach can be more effective.

Satellite imagery provides wide-area monitoring.

Crewed-aircraft LiDAR may cover entire regions.

Long-endurance fixed-wing or VTOL drones can survey larger forestry compartments.

Multirotors can collect highly detailed information over field plots and selected management areas.

Ground crews provide the calibration and validation measurements.

These datasets can then be connected.

Satellite information identifies broad patterns.

Drone data provides much finer structural information.

Field plots provide physical measurements.

This combination can create a more scalable biomass-monitoring system.

Survey Repeatability and Data Quality

Biomass monitoring depends heavily on consistency.

If two surveys are being compared, differences in sensor, altitude, season, positioning and processing can influence the result.

Survey protocols should therefore be standardised where possible.

Seasonality is particularly important in deciduous forests.

Leaf-on and leaf-off conditions can produce substantially different datasets.

LiDAR ground-return density may also change depending on vegetation conditions.

Coordinate systems and georeferencing must remain consistent.

Quality-control information should be retained with the dataset.

This allows future analysts to understand how each biomass map was produced.

Benefits, Challenges and Limitations

The main advantage of drones is their ability to connect detailed field measurements with spatial information across much larger areas.

LiDAR provides valuable three-dimensional forest structure.

Photogrammetry can create detailed canopy models.

Multispectral sensors add vegetation information.

AI can assist with individual-tree detection and biomass modelling.

GIS converts the results into practical forest-management information.

Repeat surveys can monitor growth and disturbance.

However, biomass estimation remains an indirect measurement process.

A drone does not measure tree mass directly.

Dense forests can hide understory vegetation.

Individual-tree detection may miss suppressed trees.

Photogrammetry can struggle to determine ground elevation beneath closed canopy.

Multispectral indices can saturate.

Models may not transfer reliably between different forest types.

Below-ground biomass cannot normally be directly observed.

For these reasons, field calibration, validation and appropriate uncertainty reporting are essential.

The Future of Drone-Based Biomass Estimation

Forestry biomass monitoring is likely to become increasingly integrated and automated.

Satellite data will provide frequent large-area observations.

Drone LiDAR will provide high-resolution structural information.

RGB photogrammetry may provide lower-cost repeat canopy surveys.

Multispectral and hyperspectral sensors will contribute additional vegetation information.

AI will combine these datasets with field measurements.

Forest inventory systems may automatically update biomass maps when new drone data becomes available.

Permanent sample plots could provide long-term calibration.

A forestry manager might eventually select any compartment and view estimated biomass, canopy height, species information, historical growth and carbon estimates.

Changes caused by harvesting, storms or wildfire could be incorporated into the model.

The long-term direction is toward an integrated forest-inventory platform in which field plots provide physical measurements, drones provide detailed three-dimensional forest information, satellites provide large-area monitoring, AI estimates spatial biomass patterns, GIS manages the results, and forestry professionals validate and interpret those estimates.

Conclusion

Biomass estimation is one of the most valuable analytical applications for forestry drones because it connects remote sensing directly with forest inventory, growth monitoring and carbon management.

The drone does not weigh the forest.

Instead, it measures characteristics that are related to biomass.

LiDAR provides detailed information about canopy height and vertical forest structure. Photogrammetry can create canopy models. Multispectral imagery contributes vegetation information, while field plots provide the physical measurements required to calibrate and validate the models.

The resulting biomass estimates can then be mapped across the forest.

This reveals spatial variation that traditional sample plots alone may not show.

Repeated surveys can also provide information about growth, harvesting and disturbance.

The strongest approach is therefore not drone versus traditional forest inventory.

It is the integration of both.

Ground forestry provides accurate physical measurements. Drones extend those measurements across the landscape. Satellites extend monitoring across even larger regions. AI helps process the resulting datasets, while GIS turns them into practical management information.

Together, these technologies can help forestry organisations build more detailed inventories, understand forest growth, monitor disturbance, support carbon assessments and create a continuously improving digital picture of how biomass is distributed and changing across the forest estate.