HOW SPATIAL RESOLUTION AFFECTS THE DETECTION OF TREE MORTALITY

Spatial resolution in forest health monitoring

Changes in forests often begin very locally. First, you may see individual weakened trees or small patches, and only later does the change expand enough to stand out clearly at a broader scale. That is why, in monitoring forest health, what matters is not only how often observations are collected, but also what size phenomena the data can actually resolve.

Spatial resolution refers to the sensor’s ground sampling distance (GSD), which determines the minimum size of objects or patterns that can be resolved. When the resolution is coarse, a single pixel inevitably contains multiple trees, as well as parts of the background, such as shadows, bare ground, or understory vegetation. In that case, the signal is averaged, and small-scale changes can easily mix into the overall picture. With high-resolution data, the same area is divided into more, smaller observation units. This reduces averaging and improves the ability to detect early anomalies when change is still visible only in individual trees or small groups of trees.

Forest damage and the spread of tree mortality are typically processes that unfold over time. When the goal is to track the development of tree mortality and target interventions, you need information on where change begins, where mortality is currently concentrated, and in which direction the situation is evolving. This is where high spatial resolution delivers practical value: change can be pinpointed more precisely, and monitoring can be carried out in a way that is actionable both in the field and in decision-making.

Spatial resolution refers to the sensor’s ground sampling distance (GSD), which determines the minimum size of objects or patterns that can be resolved.


Our approach leverages high-resolution multi-band imagery (red, green, blue, and near-infrared), because the Forest Health analysis focuses on changes in forest health and vitality as well as monitoring tree mortality. The near-infrared (NIR) band is essential in vegetation analysis, as it is highly sensitive to changes in chlorophyll and leaf (cellular) structure, and it complements the information derived from the visible bands.
NIR also enables false-color composites in which vegetation reflectance is highlighted differently than in natural-color imagery. Healthy vegetation typically reflects more in the NIR range than stressed or dead vegetation, which is why dead trees often stand out more clearly in analyses that incorporate NIR.
This combination is particularly well-suited for monitoring changes in forest vitality when the goal is to track tree mortality. Thanks to the high spatial resolution, we can produce data at the level of individual trees, which is essential when changes start at small scales and progress over time.

Individual-tree accuracy: what does it enable?

When data is generated at the individual-tree level, monitoring can progress from a general assessment of decline to a more detailed and operational situational overview:

  • The volumes of dead trees can be estimated accurately, for example, by a stand compartment
  • Changes can be identified and delineated more precisely
  • Development can be monitored repeatedly at the same scale
  • Decision-making can be based on targeted observations
  • When needed, detailed data can be aggregated into area-level information

Within an area of roughly 5 hectares, several individual dead trees (shown in green) can be identified in the high-resolution (0.5 m) imagery, whereas in the coarser-resolution (10 m) imagery, there are practically no pixels that can be classified as dead trees.

In forest health monitoring, resolution determines what changes can be detected, when they can be detected, and how actionable the information is for follow-up measures. When the goal is to track how the situation develops and understand the spread of tree mortality, high-resolution data and individual-tree accuracy are key.

Additionally, spatially dispersed individual dead trees can serve as indicators of biodiversity-relevant features, such as a deadwood continuum or a structurally diverse forest. When such sites are identified from high-resolution data, they can be prioritized for field verification and considered as potential conservation areas or sites of high value for nature management.

Scroll to top