Monitoring and predicting the impacts of trees on urban stormwater volume reduction

Science Center Objects

Much has been learned about how effectively individual green infrastructure practices can reduce stormwater volume, however, the role of urban trees in stormwater detention is poorly understood. This study will quantify the effect of tree removal on the urban hydrologic cycle and measure the impact that trees have on stormwater runoff volume.


The use of green infrastructure in urban areas has increasingly been accepted as a means to reduce adverse impacts of stormwater on sewer systems that service Great Lakes communities, and to a larger degree, Great Lakes water quality. Over the last decade, much has been learned about how effectively individual green infrastructure practices can reduce stormwater volume; however, one element that remains poorly understood is the role of urban trees in stormwater detention. Research is needed to provide quantitative data on how trees affect the urban water cycle through their unique influence on storage and losses such as throughfall, evapotranspiration and subsurface flows, in particular at the different spatial scales and complex urban mosaics in which bioretention and street trees are planted. While there is a body of evidence to suggest trees are an integral part of urban watershed management, few field studies have quantified the stormwater volume reduction capabilities of urban trees, and in the context of the whole-water cycle. A synthesis of past research by Kuehler et al. (2016) and Berland et al. (2017) revealed that several knowledge gaps remain on the role of trees in stormwater management. Many of these gaps center around the need for a more holistic understanding of urban tree canopy, understory, and subsurface effects on urban water cycles including, but not limited to: variation attributed to tree species, age, seasonality (leaf-on, leaf-off), regional climate, interactions with the immediate urban-suburban ecosystem, and the role of trees in the context of other green infrastructure practices (e.g., trees planted in rain gardens). These knowledge gaps apply to trees throughout the urban ecosystem (in urban forest patches, private residential and commercial parcels, as well as the public right of way), not only because of needs regarding urban water budgets, but also with respect to how these flows influence ecohydrological functions that relate to vegetation, public health, the removal of pollutants in the air and water cycle, among other ecosystem services.


The U.S. Geological Survey (USGS), U.S. Forest Service (USFS), Environmental Protection Agency - Office of Research and Development (USEPA-ORD), and the University of Wisconsin - Madison (UW) propose a field scientific investigation that will begin to fill in some of the data gaps identified by Kuehler et al. (2016) and Berland et al. (2017). The primary objective of the study is to quantify the effect of tree removal on the urban hydrologic cycle, in order to measure the impact that trees have on stormwater runoff detention volume. The study will use a paired catchment statistical design and analysis of a continuum of storm event hydrographs and additional monitoring data (e.g soil moisture and groundwater level), that characterizes how water cycles with and without trees.

This study will characterize the impact of street tree removal on stormwater runoff characteristics from two medium-density residential catchments in Fond du Lac, Wisconsin (figure 1). Catchments are comparable in size, and consist of single-family housing with approximately 0.25-acre parcels with trees interspersed along the right-of-way that flanks each street. These neighborhoods have separate septic-storm sewer systems. Stormwater runoff is conveyed via curb and gutter to stormwater inlets as the entry point to the centralized storm sewer collection system.

Map showing the GLRI urban-tree stormwater-reduction study location and catchments

Figure 1. GLRI urban-tree stormwater-reduction study location and catchments in Fond du Lac, Wisconsin.

Table detailing the physical attributes of the Control and Test catchments in the GLRI urban-tree stormwater-reduction study

Table 1. Physical characteristics of the Control and Treatment catchments, Fond du Lac, Wisconsin.

Using the paired-catchment study design, one catchment will be designated the ‘Control’ and the second catchment the ‘Treatment’ (table 1). Trees in the area, both street trees and landscape trees, appear to be similar in age and were likely planted during the time of housing construction in the late 1980s.

Table detailing the species abundance and approximate canopy cover in the GLRI urban-tree stormwater-reduction study

Table 2. Species abundance and approximate canopy cover provided by the trees in the Control and Treatment catchments.

The primary abundance of tree species lining the streets of each study catchment include ash, maple, and honey locust (table 2). Approximately one-half of the street trees in the Treatment catchment are Green Ash (Fraxinus Americana), a species of tree that is subject to disease through infestation by Emerald ash borer (Agrilus Planipennis). As such, the city of Fond du Lac plans to remove all ash trees before infestation can occur. As of October 2017, all ash trees in the Control and Treatment catchments appear to be in good health. The removal of ash trees prior to infestation provides a unique scenario to quantify the influence that street trees can have on stormwater runoff volume. The project team will establish a baseline water budget by measuring the hydrologic response of each catchment before ash trees are removed. This process will be done through evaluation of storm event hydrographs, and set the basis for comparison in the paired-catchment experimental design. Monitoring during this calibration period will span approximately 2 years, beginning in early spring 2018 before leaf-on and continue through the fall of 2019. During the winter of 2019/2020, approximately 60 percent of ash trees lining the street in the Treatment catchment will be removed, at which point a new hydrologic response will likely emerge. This change in hydrologic response would be attributed to the sudden absence of trees, mostly through the loss of interception capability due to the loss of leaves, trunks and stumps.



Data collection for this study can be split into four principle components that characterize hydrologic inputs and outputs as part of a volumetric mass water balance: 1) climate inputs and losses, 2) surface runoff, 3) sub-surface flows, and 4) arboricultural water use-losses and irrigation.

The project team will use a paired-catchment experimental design to examine the overall impact of the tree removal treatments on the various components of the local water cycle. The project team will also analyze storm sewer pipeflow hydrographs in the context of other hydrologic data. This analysis will add depth to the paired-catchment results and impart more detail as to how calibration and treatment phase hydrology compares, and establish how water fluxes redistribute through the local hydrologic cycle.

In addition to the characterization of hydrologic processes, monitoring data will be used to verify, validate, and otherwise calibrate surface runoff and base flow simulations using the i-Tree suite of forest analysis models.  The i-Tree Hydro program computes surface runoff volume and baseflow groundwater contribution to surface water flow based on weather, soil conditions, impervious cover, and tree and other vegetation parameters.  The USFS will structure the i-Tree Hydro model to evaluate conditions in the Treatment catchment before and after tree removal scenarios, and before and after future tree planting scenarios to help assess and improve tree planting designs to reduce runoff and improve water quality. Discharge data collected during the calibration period of this study will be used to calibrate and validate i-Tree Hydro’s simulation of pre- and post-treatment conditions. Using actual data to confirm or direct modification of established i-Tree computations will improve future applications of the model regarding reporting of tree- and runoff-related benefits and other ecological services provided by vegetation. This model validation will help scientists, resource managers, and communities better understand the role vegetation management can have in stormwater management, not only in the treatment area, but also in other Great Lakes watersheds and beyond.

1. Climate inputs and losses

The USGS will install at least one weather station (CSI ET107 or equivalent) comprised of sensors required to compute potential evapotranspiration rates using the modified Penman-Monteith equation. Sensors include, but are not limited to, pyranometer, anemometer, relative humidity, air temperature, precipitation, and wind speed and direction. The weather station will be located adjacent to the storm sewer monitoring shelter in the grassed median on Hawthorne Pl. Climate data will be collected at 5-minute intervals, and potential evapotranspiration will be calculated at hourly resolution. All data will be quality-assured then displayed on the USGS National Water Information System (NWIS) website. All instruments will be calibrated annually according to manufacturer recommendations except precipitation, which will be calibrated bi-annually. Collection of climate data will be suspended during winter months to prevent damage to equipment and poor data quality resulting from freezing conditions. It is assumed the period of active data collection will be in the warm-season, which we consider March through November. Climate data from the Fond du Lac County airport (National Oceanic and Atmospheric Administration station ID 72650604840) will be used to supplement periods of missing data and model simulations.

2.  Surface runoff

The USGS will measure stormwater runoff in separate storm sewer outlets that drain the Control and Treatment catchments. The storm drainage network for the Control and Treatment catchment converge to a single junction. A monitoring shelter will be placed in the grassed median adjacent to the junction and connected via buried conduit underneath the street surface.

A Doppler-type velocity sensor will be mounted to the bottom of the 15- and 21-inch storm sewer pipes draining the Control and Treatment catchment, respectively. Using USGS best practices, instantaneous pipe discharge (1-minute) will be computed by multiplying the cross-sectional area of the pipe by the corresponding mean velocity. Storm-event runoff volumes will be computed by summing the 1-minute-interval instantaneous discharge over the storm hydrograph. Because the velocity sensor only works when the depth of water is greater than 1 inch, estimates of discharge for water depths below this level will be made using Manning’s equation. A bubbler flow metering unit will be installed to verify water level and quality assure the velocity sensor’s pressure transducer data, and quantify water levels less than 1-inch. In addition to discharge, measurement of water temperature will also be made at 1-minute intervals during periods of runoff. Calibration of water level sensors will be done approximately once every month in the warm-season between March and November. Monitoring during the winter months will be done at the discretion of the principal investigator.
Retrieval of data will be done remotely by use of cellular modem, compiled, quality assured and analyzed by the USGS and USEPA-ORD. All runoff data will be stored in the USGS database and posted on the USGS NWIS website as it becomes available.

Data characterizing other hydrologic fluxes will be collected throughout the year to document changing conditions to the surrounding landscape and trees and identify potential unmeasured inputs or losses. The aggregated team will provide observations and other qualitative information that cannot be quantified, but still useful to informing on project outcomes. Photo-documentation will be employed during field visits, in addition to footage gathered by a web camera programmed to take daily photos, which will be useful in understanding runoff formation and canopy throughfall processes.

3. Tree condition, surface and sub-surface hydrologic processes

Throughout the study, the USEPA-ORD will perform near-surface soil hydrologic assessments on three occasions. A near-surface hydrologic assessment includes: taxonomy of deep soil cores (12 feet or refusal, determine soil texture of each horizon); tension infiltrometry (-2 centimeter hydraulic head) to measure infiltration rate; borehole (sub-surface) saturated hydraulic conductivity as a proxy for drainage rate just above the first hydraulically-restrictive soil horizon; dynamic penetrometry (soil mechanical resistance); and assessment of shade at each sampling position by handheld spherical densiometry). The first assessment will be done during the baseline-calibration phase in which all trees remain intact. The USGS will extract soil cores by use of a Geoprobe at 10 locations throughout the study catchments, and each will represent major land uses and/or source areas (right-of-way, lawns, etc.). Each soil core will be read according to standard US Department of Agriculture taxonomy. Resulting data will be used to gather insight on legacy hydrologic (water table) conditions. Once the soil cores have been extracted, the USGS will install piezometers and equip with instrumentation (e.g., logging pressure transducers) to measure changes to the water table at 15-minute time intervals.

In addition to soil core extraction and piezometer installation, the USEPA-ORD will conduct a series of surface and sub-surface infiltration tests at select locations throughout the study catchments. Once trees have been removed from the Treatment catchment, and the new landscape has weathered, the USEPA-ORD will replicate the previously described soil hydrologic measurements in a second assessment campaign. Removal of both trees and stumps are assumed to have a dramatic impact on surface hydrology in the Treatment catchment. A third (final) assessment will be done at the conclusion of the study (2020) to document changes in surface and sub-surface soil hydrology due to tree removal.

4. Arboricultural water use and contributions to local water cycle

The USFS, through a joint venture agreement with the UW, will be responsible for performing measurements of hydrological processes of trees. Specific tasks include:

  • Throughflow measurements: The urban canopy intercepts precipitation, reducing and spatially redistributing the effective precipitation that reaches the land surface, which could otherwise become stormwater runoff
    • Continuous estimates of interception from storms will be obtained by comparing rainfall records collected beneath the canopy with sensors in open areas. Six rain gauges will be installed beneath the canopy and monitored for the study period.
    • The spatial heterogeneity of throughflow will be determined for two storms per year. Precipitation totals will be collected using a dense network of precipitation gauges (~200) beneath the urban canopy on both control and treatment blocks. The network will be based on transects along the street terrace both at the curb and the sidewalk.
  • Sapflow measurements:  Three trees will be instrumented in triplicate with sap flow sensors and data will be collected during the growing season.
  • Soil moisture measurements:  Soil moisture sensors will be installed at depths of 4, 14, 24, 44, 64, and 84 centimeters at a distance of less than 2 meters from the trees where sapflow is monitored.
  • Infiltration measurements:  Infiltration dynamics will be inferred from the wetting fronts observed at each soil moisture monitoring station.  In addition, infiltration capacity will be measured at select locations distributed across the study site.
  • Tree characteristics:  GIS analysis will be performed to calculate the crown area of each tree in the study area. For each tree, the area over pervious and impervious areas will be calculated. The diameter at breast height will be measured for each tree on the street terrace so that sap flow estimates can be scaled up. Optical estimates of leaf area index will be measured monthly during the growing season on transects along the street terrace both at the curb and the sidewalk. Phenology will be monitored with cameras.

The soil, tree and meteorological measurements will be evaluated against the backdrop of storm flows, meteorological data, and soils data, with particular attention to how hydrologic processes change when trees are removed in the Treatment catchment. This information will be used in various modeling projects examining the effects of trees on hydrological processes.



There will be three products summarizing the effect of urban trees on stormwater volume reduction. The first product will be a report characterizing the hydrologic response of the Control and Treatment catchments based on current conditions during the calibration phase. The intent of the report will focus on seasonal variation of surface and sub-surface hydrologic measurements as they relate to existing tree canopy density and species. The second product will be a final report summarizing changes in surface and sub-surface hydrology in the Treatment catchment after tree removal. The format of each report will be a USGS Scientific Investigation Report or a journal article with inclusive supporting data tables. The third product will be incorporation of field-level data into the i-Tree model to provide improved hydrologic modeling capability for urban trees. Additionally, a series of ‘town-hall’ style meetings will be held near the study area to inform both residents and public officials on the progress of the study as well as any conclusions that can be made based on collected data.



Chart showing the estimated timeline of major milestones in the GLRI urban-tree stormwater-reduction study

Estimated timeline of major study milestones


Kuehler E, Hathaway J, Tirpak, A. Quantifying the benefits of urban forest systems as a component of the green infrastructure stormwater treatment network. Ecohydrology. 2017;10:e1813.

Berland, A., Shiflett, S.A., Shuster, W.D., Garmestani, A.S., Goddard, H.C., Herrmann, D.L. and Hopton, M.E., 2017. The role of trees in urban stormwater management, Landscape and Urban Planning, 162, pp. 167 – 177.