Kamloops BC

Resource or Hazard? Design Implications for Stemflow from Urban Trees

We’ve all taken shelter under a tree’s canopy during a storm, grateful for its ability to intercept rain. You may not realize this, but rainwater that passes through tree canopies is modified both volumetrically and chemically in processes that have been relatively well studied in forests. That modification, which encompasses several different components, is what makes being under tree a good place to stay dry.

The few studies done in urban environments suggest that hydrological processes in isolated trees, such as those in streets or parking lots, differ from those in forests – yet many models and calculations of stormwater-related benefits assume that the processes and quantities are similar. Growing recognition of the many values provided by urban forests has stimulated investment in green infrastructure and intensified the need for studies specific to urban trees­. We aimed to test the assumption that forest trees and urban trees perform similarly for one specific component of understory precipitation, stemflow, in a manicured urban park.

What is Stemflow?

Figure 1. Stemflow in the context of other water balance processes. (Source: Inkiläinen et al., 2013, adapted from Levia and Frost, 2006.)Figure 2. Aerial view of McArthur Island Park.

Figure 1. Stemflow in the context of other water balance processes. (Source: Inkiläinen et al., 2013, adapted from Levia and Frost, 2006.)

Of the total rain that falls on a tree canopy (Figure 1):

  • Stemflow is the portion that is funneled to and concentrated at the base of the trunk—in deciduous forests, this typically represents less than 3  percent of the total rainfall, but may not be negligible for isolated urban trees!
  • Throughfall is the portion that reaches the ground diffusely, passing through gaps in the canopy, or drips from leaves, twigs, and branches comprising the canopy.
  • Interception loss is the portion of rain that remains stored in the canopy to be evaporated, never reaching the ground.

To reduce flooding caused by stormwater, urban foresters, landscape architects, and stormwater engineers have typically aimed to maximize interception loss and minimize understory precipitation. But what if this approach has been misinformed? We wanted to examine the assumptions behind it.

What if stemflow is not negligible for some commonly planted trees?

Could informed tree selection and planting allow us to manage stemflow as a resource, complementing how we are starting to manage stormwater runoff from surfaces?

Could designs for stemflow infiltration also accommodate integrated stormwater management practices?


Figure 23. Aerial view of McArthur Island Park.

Figure 23. Aerial view of McArthur Island Park.

Our study site was an intensively managed urban park in semi-arid Kamloops, BC, Canada (Figure 2). Stemflow and weather data were collected between June, 2012 and November, 2013 using the following methods:

  • Stemflow was collected by a system of collars and reservoirs (Figure 3)
Figure 3: Figure 3. Stemflow collection system.

Figure 3: Figure 3. Stemflow collection system.

  • Our sample included 37 isolated deciduous trees representing 21 commonly specified cultivated species (e.g., ash, maple, linden, and beech).
  • The 37 specimens had a range of sizes (10–69 cm [4–27”] diameter at breast height, or DBH) and diverse traits; single- and multi-leader tree groups were analyzed separately.
  • A weather station collected data for 86 rain events over 17 months, more than 80% of which were less than 5 mm (0.2”) in depth.

Highlights of our Findings

There were four main highlights of our findings.

For single-leader trees, we found that higher stemflow rates were associated with large trees, low branch count, large leaf size, and high-relief bark (particularly when linearly furrowed).

For multi-leader trees, we found that higher stemflow rates were associated with plentiful leaders, high wood cover and branch angles, narrow canopies, and smooth bark.

For smaller trees, stemflow was initiated at lower rain depth for those with smoother bark and high branch angles on single-leader trees, and with more leaders on multi-leader trees.

Lastly, for all trees, stemflow tended to increase with wind speed and rain inclination angle, and decrease with higher vapour pressure deficits (as found by other researchers).

Record Holders

We had two record-holding specimens in our study that outperformed all the other trees.

The first was a Columnar English Oak (see branch structure in Figure 4). This tree funnelled 22.8% of the water from a 25.6 mm (1.0”) rain event, and had an impressive 11.8 ± 9.1% average over the study period. Why was this the case? We believe it’s because this tree had the highest branch angles and canopy height-to-width ratio in the study, as well as very dense wood cover within its canopy.

Figure 4. Leafless sketch of Columnar English Oak.

Figure 4. Leafless sketch of Columnar English Oak.

The second record holder was an American Beech (Figure 5). For a 25.6 mm (1.0 “) rain event, stemflow concentrated at the base of this tree was 197 times deeper than rain depth beyond the canopy (known as a “funnelling ratio”, a measure of a canopy’s stemflow-producing efficiency whereby a ratio of 1 means depth at the base equals rain depth). Why? This tree had many leaders, high upper-branch angles and wood cover, and smooth bark (possibly hydrophobic, which would increase the flow of water even more).

Figure 5. Leafless sketch of American Beech.

Figure 5. Leafless sketch of American Beech.

Simply put, stemflow was certainly not negligible for these or many other study trees!

Design Implications

Most importantly, we showed that stemflow is not necessarily negligible, as had previously been assumed, particularly for certain tree species, forms, and sizes.

For existing streetscapes, this means a few things. If soils are highly compacted and/or paving dominates the site, as is the case in most urban areas, it is unlikely that either stemflow or other stormwater is infiltrating at the base of the tree. The risk of concentrated stemflow contributing to runoff quantity and quality issues depends on the rainfall regime, airborne pollutant conditions, size and traits of existing trees, and consequences of mismanagement (e.g., sensitive ecosystems). Since soil compaction is challenging to mitigate, controlling runoff through various best management practices (e.g., raingardens or rock pits/trenches) is recommended in this situation.

For proposed streetscapes, or other areas where paving is prevalent such as parking lots or plazas, there is more opportunity to apply our findings to site design.

Ideally, designs will ensure infiltration capacity by specifying appropriate soils and generous soil volumes. If stemflow can infiltrate into either grass or permeable soil at the base of the tree (such as into a suspended pavement system, open planter, or raingarden) for the majority of a location’s anticipated storm events, then species may be chosen that promote stemflow. The benefits of infiltrating stemflow can be significant, and include: 1) reduction of runoff from water quantity and quality perspectives; 2) self-irrigation, with the need for supplementation depending on the species and climate; 3) self-nourishment with nutrient-rich stemflow; 4) biofiltration of pollutants washed from the canopy into soils via stemflow; and 5) groundwater recharge.

If infiltration into surrounding soil is limited, then trees may be selected for non-conducive traits. However, reducing stemflow may result in increased throughfall, which is a more difficult quantity of rainwater to manage as it typically falls diffusely on pavement beneath the canopy.

Interception loss has its limits; if a greater percentage of drainage from the canopy is in the form of stemflow (concentrated) rather than throughfall (diffuse), there is a greater opportunity to manage this input (e.g., infiltration into soil at base of tree).


Can “green infrastructure” accommodate stemflow while meeting many other objectives? Definitely.

Various articles from this blog describe compatible strategies, including directing stormwater into Silva Cells, using trees and soils to manage stormwater, and the role of trees and plants in bioretention. Other components in the site-scale water balance include evapotranspiration, which can return significant volumes of stormwater from soils beneath trees to the atmosphere, and groundwater recharge whereby stemflow has been shown to follow preferential pathways along roots.

While more systematic urban-tree studies are needed in various climates, we have confirmed the importance of high branch angles, bark relief, and wood cover for stemflow.

Novel findings of our research included the stemflow-promoting role of linearly furrowed bark in single-leader trees’ stemflow rates and the association of stemflow production with different traits for single- vs. multi-leader canopy structures. Further study is needed on these topics as well as on urban conifers and stemflow chemistry, but the key  take-home messages from this study are:

  1. Recognize urban trees as part of the hydrological system. Stemflow can be a valuable input (or problematic in excess), so tree selection, siting, and planting design should reflect a tree’s anticipated contribution to site hydrology over its lifespan.
  2. Provide sufficient quality and quantity of soils to absorb and biofilter stemflow (as well as throughfall and runoff) and support growth of trees to maturity.
  3. Integrate trees with broader infrastructure designed to manage rainwater at the source.

About the Researchers

Julie Taylor Schooling, MScES, MBCSLA, is a Landscape Architect with McElhanney Consulting Services Ltd. who has long had an interest in the integration of stormwater management and landscape features. You can contact her at [email protected]. Darryl Carlyle-Moses, PhD, is Associate Professor and Chair in the Department of Geography & Environmental Studies at Thompson Rivers University in Kamloops, BC. His forest hydrology expertise and Julie’s passion for urban trees helped to frame and answer this study’s research questions. You can contact him at [email protected].

Schooling, J.T. 2014. The influence of tree traits and storm event characteristics on stemflow production from isolated deciduous trees in an urban park. MScES Thesis. Faculty of Science, Thompson Rivers University: Kamloops, BC. 103 pp. URL: www.kamloops.ca/stormwatertrees

David WiseCC BY 2.0

Stormwater Trees logo block


Leave Your Comment

Join Our Newsletter

Sign up to our newsletter to stay up-to-date on news, updates, and product information.

Explore our archives

Interested in a DeepRoot product? Get a quote today.

Request a Quote [email protected]

Tel: 415 781 9700
Toll Free: 800 458 7668
Toll Free: 800 277 7668
Fax: 415 781 0191

Sign up to our newsletter to stay up-to-date on news, updates, and product information.