The research examined here does not establish a universally safe root diameter, distance from the trunk, number of roots, or percentage of root loss. A tree may remain alive after root cutting while experiencing reduced growth, and the study does not resolve questions about decay, structural stability, or effects beyond its observation period. If proposed work could affect a substantial part of a tree’s root system—or if tree failure could harm people or property—do not authorize the cut from a generic threshold. Pause irreversible work and obtain a site-specific assessment.
Related: How Palm Tree Roots Grow: Structure, Depth and Spread.
The short answer: do not rely on a universal safe-cutting rule
The available Douglas-fir study cannot tell a homeowner how much of a particular living tree’s root system may be cut safely. Its findings concern a defined experimental population and do not provide a formula based on visible root diameter, root count, distance from the trunk, or an estimated percentage of root loss.
Survival is also too narrow a definition of safety. In the study, greater root-damage severity was associated with lower estimated maximum diameter and height even though induced root damage did not have a statistically significant effect on cumulative survival. The research did not quantify structural stability or decay outcomes.
Generic percentages and distance rules therefore should not be used to approve landscaping, trenching, paving, or construction around a particular tree. Factors such as species, maturity, soil, existing condition, root arrangement, and the form of disturbance may differ from the experimental conditions; this study does not determine how those differences alter the outcome.
When substantial root disturbance is possible or the consequences of failure would be serious, an on-site assessment is the conservative course. Such an assessment can account for the actual tree, site, existing damage, and proposed work, but it cannot guarantee that cutting a particular root will be harmless.
What the 10-year Douglas-fir study actually found
Dryw A. Jones, Constance A. Harrington, and David Marshall followed 1,440 trees for 10 years across six young, even-aged Douglas-fir plantation stands in Capitol State Forest near Olympia, Washington. Researchers exposed roots near the stems and severed selected smaller roots or girdled selected larger roots in replicated damage and control treatments. These methods and findings are described in their study, “Survival, and Growth Response of Douglas-Fir Trees to Increasing Levels of Bole, Root, and Crown Damage”.
Under those experimental conditions, induced root damage did not have a statistically significant effect on cumulative survival. Increasing root-damage severity was nevertheless associated with lower estimated maximum diameter and height. Increasing damage across roots, bole, and crown was associated with growth reductions, while bole damage was a stronger predictor of mortality than root damage under the study conditions (Jones, Harrington, and Marshall, “Survival, and Growth Response of Douglas-Fir Trees”).
That distinction is essential: an absence of a statistically significant survival effect is not an absence of biological effects. It also does not establish that an injured tree remained structurally stable or free from decay.
The study additionally illustrates why root injury should not automatically be considered in isolation. Trunk and crown damage were incorporated into the analysis, including natural crown damage that occurred during the study. A residential tree with existing injuries is not equivalent to an otherwise comparable tree receiving an isolated experimental root treatment.
The defensible conclusion is narrow. The experiment provides evidence about survival and growth among young Douglas-fir in six plantation stands over 10 years. It does not provide general instructions for cutting roots on mature residential trees, other species, or trees exposed to different forms of excavation damage.
Why the study’s 40% figure is not a homeowner safety threshold
The study’s management implication described root damage below 40% of estimated total root cross-sectional area as unlikely to affect survival, while allowing for minor growth reductions. “Unlikely to affect survival” is not a guarantee of survival, unchanged health, structural stability, or freedom from decay (Jones, Harrington, and Marshall, “Survival, and Growth Response of Douglas-Fir Trees”).
Cross-sectional root area refers to the combined cross-sectional area represented by damaged roots relative to an estimate of the entire root system. The researchers estimated that whole-system relationship rather than directly measuring every root. The figure does not mean:
- 40% of the visible roots
- 40% of roots counted individually
- The sum of visible root diameters
- 40% of the roots exposed on one side
- Roots beyond a particular distance from the trunk
Counting exposed roots or adding their diameters would not reproduce the study’s measurement. Nor would looking only at roots revealed by a trench, because those visible roots do not represent a direct inventory of the complete root system.
The achieved damage levels also did not divide neatly at a single threshold. Recalculated damage ranged from 15% to 23% in the nominal 25% treatment and from 31% to 45% in the nominal 50% treatment (Jones, Harrington, and Marshall, “Survival, and Growth Response of Douglas-Fir Trees”). Because the higher treatment spanned values on both sides of 40%, the experiment does not establish 40% as a precise biological cutoff.
The figure therefore cannot be converted into permission to cut “less than 40%” of a residential tree’s roots. Even if the affected share could be estimated accurately, the experiment did not establish unchanged growth, structural safety, absence of decay, or outcomes after the 10-year study period.
Use this evidence-boundary table before making a decision
| Question | What the evidence supports | What to do with the uncertainty |
|---|---|---|
| Will the tree immediately or eventually die? | Induced root damage did not significantly affect cumulative survival over 10 years in the studied young Douglas-fir (Jones, Harrington, and Marshall). | Do not assume that result transfers to a different tree, species, site, or injury pattern. |
| Will growth remain unchanged? | Greater root damage was associated with lower estimated maximum diameter and height. | Treat survival and unchanged growth as separate outcomes. |
| Is less than 40% safe? | The figure concerns estimated total root cross-sectional area in this experiment (Jones, Harrington, and Marshall). | Do not use it as a universal cutting allowance or precise cutoff. |
| Can I judge risk from root diameter or root count? | Those measures do not reproduce the study’s whole-system estimate. | Do not authorize cutting from either measure alone. |
| Will the tree remain stable or avoid decay? | The study does not quantify those outcomes. | Keep those uncertainties separate from the survival finding. |
| Does this apply to my mature residential tree? | The study involved young Douglas-fir plantation stands and controlled treatments. | Recognize that species, maturity, soil, climate, condition, and injury method may differ materially. |
The less closely a proposed cut resembles the experiment—and the greater the consequences if the tree fails—the less defensible it is to extrapolate from the results.
When not to make the cut based on this evidence alone
The study does not validate a do-it-yourself root-cutting procedure or explain how to identify which individual roots can be removed safely. It should not be used by itself to approve work involving:
- A mature residential tree
- A species other than Douglas-fir
- Multiple roots or extensive excavation
- A tree close to a home, road, play area, parking space, or another occupied location
- A tree with existing trunk, crown, or root damage
- Root cutting combined with trenching, tearing, compaction, paving, or grade changes
These circumstances are not presented as diagnostic rules. They are reasons the Douglas-fir experiment may provide an inadequate basis for a particular decision and why the consequences of extrapolating incorrectly may be substantial.
Construction disturbance can also differ materially from the controlled treatments. The experiment involved selected lateral roots near the stem being severed or girdled. A project may disturb roots and surrounding soil in several ways at once, so its effects cannot be equated automatically with a selected experimental cut.
Existing trunk and crown damage likewise should not be set aside when considering the study. Damage across roots, bole, and crown was associated with growth outcomes, and bole damage predicted mortality more strongly than root damage under the experimental conditions. The practical question is not merely whether one root can be cut, but whether this research adequately represents the tree’s cumulative condition and the proposed disturbance.
If roots have already been severed, pause additional irreversible disturbance rather than using the study’s survival finding to justify more cutting. Where biological uncertainty and potential consequences are substantial, arrange an on-site assessment of the actual tree and site. The study supports caution—not a universal percentage, diameter, distance, or root-count rule.