ResearchTechnology and Digital PracticeClimate Impact Simulation Tools

Microclimate and stormwater simulation tools

What ENVI-met, Ladybug Tools, SWMM, the National Stormwater Calculator and i-Tree compute, what they need, and how far to trust their results in design.

Field Method · AI-assisted research roundupPublished Sources checked

A simulation tool turns a site model and a weather record into numbers. Every result has two parts: what the software computes from physics, and what it assumes because you did not supply it or it cannot represent it. Whether an output can carry a design decision depends on which part dominates, and on whether anyone has compared that tool with measurements.

Below we sort the tools a landscape architect is likely to meet into four classes: sun and shade, outdoor thermal comfort, stormwater and tree canopy. For each we record what the maker’s documentation says is computed, what must be supplied, and what published comparisons show. This is not engineering advice. We ran none of these tools, so every statement about behaviour comes from documentation or papers.

Sun and shade

The Ladybug Tools SunPath component draws a sun path and computes sun positions for hours of the year from a location, a north angle and an optional daylight-saving period. It also outputs sun vectors intended for solar access and shading design. Ladybug Primer, SunPath The Ladybug source code says its sun position routine was originally written by Trygve Wastvedt, based on NOAA’s method, and modified by the Ladybug developers. Neither the source file nor the primer page states a positional accuracy. Ladybug source, sunpath.py

Sun position depends only on place and time. Our reading: a shade study is only as good as the geometry that casts the shade. In a park that is mostly trees, so canopy form, season and modelled canopy porosity decide the result, and the SunPath page does not address them. Shade also does not remove the heat a person exchanges with nearby surfaces, which comfort tools model.

Outdoor thermal comfort

Regional heat screening, covered in Climate risk screening tools for site design, tells you whether heat is a concern for the area. The tools below compare design responses at the scale of a plaza. Both report mean radiant temperature (MRT) and indices built on it, such as the Universal Thermal Climate Index (UTCI).

ENVI-met. The help centre describes a three-dimensional, non-hydrostatic microscale model of surface, plant and air interactions, with horizontal resolution typically 1 to 10 m, simulation periods commonly 24 to 48 hours and time steps of 1 to 5 seconds. Computed processes include radiation with shading and reflection, vegetation-air exchange, façade and roof temperatures, and soil water and heat transfer. Its BioMet module computes MRT, PMV/PPD, PET and UTCI. ENVI-met help, model architecture overview The atmospheric model solves Reynolds-averaged Navier-Stokes equations on a grid with an E-epsilon turbulence closure. The page says that closure often fails in low wind and suggests a mixing-length model there. ENVI-met help, atmospheric model Plants are three-dimensional and described by leaf area density, which governs how much sun they intercept, and the model returns latent and sensible heat from them to the air. ENVI-met help, vegetation model The pages give no leaf area density values, so a run’s plant parameters are your or the library’s assumption.

The help centre lists version 6.0, dated September 23, 2026, as the current release. Its notes list improved longwave radiation physics, a breakdown of pedestrian MRT into solar, reflected and longwave parts, and a spin-up that sets initial soil temperatures in place of manual input. They mark Fast Thermal Comfort and Wind Comfort as exclusive to the Business licence tier. ENVI-met help, Version 6.0 None of the comparisons below tested version 6.0. A paper that compared the program with a Ladybug Tools workflow states that full forcing of meteorological inputs, from a weather file, became possible from version 4.4.0. Ibrahim, Kershaw and Shepherd 2020

The licence types page lists Commercial, Science, PhD and Student licences. Only the Commercial licence permits commercial use, and only it includes solar access simulation, wind simulation and wind comfort, and Fast UTCI calculation and statistics. The other three are non-commercial and limited to holistic simulations. The student licence is capped at a model of 100 × 100 × 50 grid cells and 50 simulations. Every licence needs a personal One Click LCA account. The page shows no prices. ENVI-met help, licence types We read the home page and five help-centre pages. None of them names a validation study, and we did not search the maker’s publication lists. ENVI-met

Ladybug Tools. The maker describes free applications that connect Rhino and Grasshopper to simulation engines including Radiance, EnergyPlus/OpenStudio and OpenFOAM. Ladybug Tools The Grasshopper plugin repository shows an AGPL-3.0 licence; we did not read the host software’s terms. ladybug-grasshopper on GitHub

The UTCI map recipe needs a Honeybee model with sensor grids and a weather file. Longwave MRT multiplies Radiance view factors by EnergyPlus surface temperatures; shortwave MRT traces rays from each sensor to the sun. The documentation states three assumptions that matter in a landscape. Outdoor air temperature and relative humidity are taken directly from the weather file. Outdoor context shades and the ground are taken to be at that air temperature unless they are modelled as Honeybee rooms. Outdoor wind comes from the weather file unless you supply per-sensor wind. The page does not describe evaporative cooling by vegetation or water. Honeybee primer, UTCI Comfort Map In this workflow, then, a tree changes MRT through its shade but does not cool the air. Our reading: where the design question is whether planting or water cools the air, the ENVI-met documentation describes that exchange and the Ladybug comfort map documentation does not.

What UTCI assumes. In the Ladybug component UTCI takes air temperature, MRT, relative humidity and wind speed at 10 m, with a default wind speed of 0.5 m/s. It assumes a person walking at about 2.4 met who adapts clothing to the temperature, and its default no-stress range is 9 to 26 °C. The page recommends PET for situations that do not fit those assumptions. Ladybug Primer, UTCI Comfort A person seated under a tree falls outside it.

What comparisons with measurements show

We found four papers that compare these tools with measurements or with each other. None tests a current release on a landscape project.

  • ENVI-met 3.1 against field data, Lisbon. Szűcs and colleagues measured MRT in a park and street canyons on four summer days in 2006 and compared it with ENVI-met 3.1, RayMan and SOLWEIG. In direct sun ENVI-met overestimated MRT by about 20 percent, which they report as 15 °C on average and up to 20 °C in some cases. In shade it had the best fit of the three. Szűcs, Gál and Andrade 2014
  • ENVI-met 5 against field data, hot Mediterranean climate. Aleksandrowicz, Saroglou and Pearlmutter compared monitored and simulated summer MRT at 60 locations to test the radiation scheme introduced in version 5. Their abstract reports that accuracy improved, but that the software systematically underestimated MRT in both shade and full sun, with RMSE of 6.08 °C in shade and 13.32 °C in full sun. The authors suggest the error may stem from how longwave emission by ground surfaces is represented. We read only the abstract. Aleksandrowicz, Saroglou and Pearlmutter 2023
  • Ladybug Tools against field data, Turin. Nicholson and colleagues compared Ladybug Tools 1.7 with rooftop measurements made on 1 July, to study lightweight shade structures. The standard workflow overestimated the shortwave MRT increment by up to 3 K, and its longwave error ran the other way and masked it. Setting Radiance ambient bounces to 6 reduced run-to-run variation but raised the overestimate by up to 7 K. With ground reflectivity lowered from 0.30 to 0.15 and ambient bounces set to 6, shortwave MRT fell within ASHRAE calibration limits (R² 0.87, cvRMSE 5.70%, nMBE −0.80%), but predicted ground surface temperature was 10 K too high. The authors judge the tools valid for early-stage design, say users should assume shortwave MRT is overestimated, and in their case study treat relative differences between options as the metric that matters. Nicholson et al. 2024
  • Ladybug Tools against ENVI-met, hypothetical layouts. Ibrahim and colleagues compared Ladybug Tools with ENVI-met v4.4.3 on three generic layouts using a Cairo weather file. They report MRT RMSE of 5.41 to 5.65 °C and UTCI RMSE of 2.42 to 2.45 °C, and describe agreement as acceptable from 8 am to 5 pm. At individual receptors the MRT difference reached 32.5 °C in the afternoon. This compares two models, not a model with measurements. Ibrahim, Kershaw and Shepherd 2020

Our interpretation: direct-sun MRT shows the largest errors, and for ENVI-met the direction has changed. Version 3.1 overestimated MRT in direct sun; version 5 underestimated it in sun and shade. Version 6.0 changes the longwave physics again, so we would not assume its error runs either way until it is compared with measurements. Comparing options is better supported than quoting a predicted comfort value, within the limits set out below.

Stormwater

SWMM. EPA describes the Storm Water Management Model as a tool for runoff quantity, quality and hydraulic routing. It lists SWMM 5.2.4, with installers dated 08/07/2023, and states that the software is public and free. It models eight generic low impact development (LID) controls, among them bioretention cells, permeable pavement, green roofs, rain gardens and vegetative swales. EPA, Storm Water Management Model

The 5.2 User’s Manual (EPA/600/R-22/030, February 2022) lists what a subcatchment needs: rain gage, outlet, area, percent imperviousness, average slope, characteristic width of overland flow, Manning’s n for pervious and impervious areas, and depression storage. Infiltration can follow Horton, Green-Ampt or Curve Number methods. The manual describes two ways to place LID controls (pp. 82–83). Several controls placed within one subcatchment act in parallel and cannot be placed in series. A subcatchment given over to a single control can receive runoff from upstream subcatchments, so controls can be strung in series. Either way, percent impervious, width and sometimes area may need adjusting after placement. SWMM 5.2 User’s Manual The National Stormwater Calculator guide says width is usually set from area and flow-path length, then refined by calibration against measured hydrographs. NSC web app user’s guide A SWMM model without gauge data is therefore uncalibrated, and the width and soil values are assumptions.

EPA also states that its SWMM climate tool, SWMM-CAT, has not been maintained since March 18, 2025, and points to the upcoming NOAA Atlas 15. EPA, Storm Water Management Model

National Stormwater Calculator (NSC). EPA describes a web tool that estimates annual rainwater amounts and runoff frequency for a site with green infrastructure. Version 3.4.0 of the web application is dated 08/10/2022, and the desktop version is marked no longer supported. EPA, National Stormwater Calculator The user’s guide says it runs the SWMM 5.1.015 engine over an hourly record, updated from 1970 to 2006 to 1990 to 2019. It is most appropriate for screening-level analysis of sites up to 12 acres with uniform soil. It infers Green-Ampt infiltration from the Hydrologic Soil Group and does not track infiltrated water afterwards. NSC web app user’s guide The EPA page gives the same 12-acre limit for hydrology analysis but, under applications, describes screening sites up to several dozen acres. It adds that the calculator does not accept user-supplied weather data, offers seven LID controls and cannot model treatment trains, for which it points users to the full SWMM model. EPA, National Stormwater Calculator Where the calculator fits among site-scale stormwater strategies is covered in Site-scale strategies for heat, stormwater and drought.

Its climate scenarios are CMIP5 projections under RCP 8.5, packaged in EPA’s CREAT 3.1, with Hot/Dry, Central and Warm/Wet groupings for 2035 and 2060. The guide advises considering all three when bracketing future conditions. NSC web app user’s guide The only validation in the guide concerns cost estimates: three of five regional case studies fell within the calculator’s range. We found no comparison of its runoff with site measurements.

Trees and canopy

i-Tree is a suite. i-Tree Eco takes field data from single trees, inventories or random plots, with local hourly weather and pollution data, and estimates carbon, avoided runoff, pollution removal and other benefits. The page describes Eco version 6 and lists forecasting with tree planting inputs. i-Tree Eco The tools page says Eco requires installation on Windows, and that i-Tree Canopy estimates land cover from random points on aerial imagery. It also lists tools for tree planting: i-Tree Planting models the future benefits of a planting project, and i-Tree Stormwater its runoff benefits over time. i-Tree Design estimates benefits of individual trees, with forecasting. i-Tree tools The methods page says Canopy, Landscape, Design, Planting, MyTree and OurTrees use benefit multipliers, or base values, built from batch runs of i-Tree Eco. It lists method documents for each tool and says Hydro development has stopped. The i-Tree pages state that a change in USDA Forest Service policy has indefinitely halted traditional funding for i-Tree, so support and updates may change. i-Tree methods and files

A peer-reviewed paper describes the Urban Forest Effects (UFORE) model and its ground-based plot sampling, with results for 14 US cities. Nowak et al. 2008 The Forest Service’s assessment guide states that 200 one-tenth acre plots typically give a relative standard error under 15 percent for the whole-population estimate. For photo-interpretation, 100 points give a standard error of about 4.6 percent and 1,000 points about 1.4 percent, assuming 30 percent canopy cover. It warns that interpreters can misclassify points. USFS, A Guide to Assessing Urban Forests The guide is undated; its file dates from 2012, and it names i-Tree tools the current tools page no longer lists.

Our interpretation: Eco and the sampling guide describe an existing tree population, with stated sampling error. The planting tools project benefits of new trees by applying multipliers derived from Eco. We did not read their method documents, so we cannot say how well those multipliers fit one new planting on one parcel.

Reading a result in a design decision

This list is our interpretation of the sources above.

  1. Prefer comparison to absolute values, and compare like with like. The Turin authors treat relative differences between options as the useful metric. A ranking is weakest when the options differ in an input the model is sensitive to, such as ground reflectivity in light against dark paving, or a tree canopy against a fabric shade. Run the sensitivity in item 4 before trusting it.
  2. Check which inputs you chose and which the software chose: ground and air temperature, wind, soil group, flow width, leaf area and the weather record. In a park shade study the main assumptions are tree geometry, season and canopy porosity.
  3. Stay inside the stated domain. The calculator’s guide is written for sites up to 12 acres with uniform soil. UTCI assumes a walking adult.
  4. Run the sensitivity that matters. The Turin study found shortwave MRT highly sensitive to ground reflectivity.
  5. Record, or ask a consultant to report, the software version, forcing method, weather file, grid resolution, leaf area density and its source, ground reflectivity, wind source, calibration status and every changed default, so a reviewer can reproduce the run.

Simulation output does not replace the responsible authority’s method. Stormwater design for a permit follows the local agency’s manual. We found no regulatory standard for thermal comfort in these sources; check the jurisdiction.

What we could not settle

  • How ENVI-met 6.0 compares with measured MRT. We found no field comparison of version 6.0, and we read only the abstract of the version 5 study.
  • Whether the Business licence tier named in the version 6.0 notes is the Commercial licence on the licence types page.
  • The positional accuracy of Ladybug’s NOAA-based sun position routine.
  • Whether any Ladybug Tools workflow beyond the comfort map adjusts air temperature for planting or water. We read only the comfort map and UTCI pages.
  • The runoff accuracy of the National Stormwater Calculator against measured sites, and which of EPA’s two site-size statements governs.
  • How i-Tree’s planting tools (Planting, Design) handle a single new planting, how the UFORE model relates to Eco version 6, and the effect of the funding halt on future versions.
  • The licensing terms of the host software for Ladybug Tools.

Using this in Field Method

Site Lens assembles site context for a San Diego County site: terrain from the 2024 county lidar survey, building footprints, lidar tree estimates and a draft surface map whose area table gives pervious, impervious and unknown totals, with the impervious total as a range. Its page lists solstice sun-hours layers computed from the site elevation grid. It does not say whether buildings or trees shade them, so they do not replace a shade study. Site Lens does not run thermal comfort or stormwater simulations. Use it to gather the geometry, trees and surface areas the tools above need, and record which values are mapped and which assumed.

Sources

  1. Ladybug Primer — SunPath component
  2. Ladybug source code — ladybug/sunpath.py
  3. ENVI-met Help Center — Model architecture overview
  4. ENVI-met Help Center — Atmospheric model
  5. ENVI-met Help Center — Vegetation model
  6. ENVI-met Help Center — ENVI-met updates - Version 6.0 - September 2026
  7. Ibrahim, Kershaw and Shepherd (2020) — A methodology for modelling microclimate: a Ladybug-tools and ENVI-met verification study
  8. ENVI-met Help Center — Available license types for ENVI-met
  9. ENVI-met — Microclimate software home page
  10. Ladybug Tools — home page
  11. ladybug-grasshopper repository
  12. Honeybee Energy Primer — UTCI Comfort Map
  13. Ladybug Primer — UTCI Comfort component
  14. Szűcs, Gál and Andrade (2014) — Comparison of measured and simulated Mean Radiant Temperature. Case study in Lisbon (Portugal)
  15. Aleksandrowicz, Saroglou and Pearlmutter (2023) — Evaluation of summer mean radiant temperature simulation in ENVI-met in a hot Mediterranean climate (Building and Environment 245, 110881)
  16. Nicholson, Nikolopoulou, Watkins, Löve and Ratti (2024) — Data driven design for urban street shading: validation and application of Ladybug Tools as a design tool for outdoor thermal comfort
  17. US EPA — Storm Water Management Model (SWMM)
  18. US EPA — Storm Water Management Model User's Manual Version 5.2 (EPA/600/R-22/030)
  19. US EPA — National Stormwater Calculator Web App User's Guide, Version 3.4.0 (EPA/600/C-22/150)
  20. US EPA — National Stormwater Calculator
  21. i-Tree — i-Tree Eco
  22. i-Tree — Tools
  23. i-Tree — Methods and Files
  24. Nowak, Crane, Stevens, Hoehn, Walton and Bond (2008) — A ground-based method of assessing urban forest structure and ecosystem services
  25. USDA Forest Service — A Guide to Assessing Urban Forests (NRS-INF-24-13)

These tools support site review and design studies. Read each tool’s stated capabilities and limits before use.

Explore Site Lens
Back to Research