Mendoza et al. (2026) Process-based evaluation of green roof models for assessment of heat mitigation efficacy in WRF (v4.3.1) and EnergyPlus (v8.6.0)

Identification

• Journal: Geoscientific model development
• Year: 2026
• Date: 2026-10-12
• Authors: Maria Martinez Mendoza, Alireza Saeedi, James Voogt, E. Scott Krayenhoff
• DOI: 10.5194/gmd-19-9729-2026

Research Groups

• Department of Geography and Environment, The University of Western Ontario, London, Ontario, Canada
• School of Environmental Sciences, University of Guelph, Guelph, Ontario, Canada

Short Summary

This study evaluates the performance of two green roof models, EcoRoof (ER) and a multilayer green roof parametrization for WRF (WRF-MLGR), against observations from a monitored green roof in London, Ontario. The results show that ER provides a more consistent representation of the daytime energy balance, while WRF-MLGR exhibits structural biases in surface heating.

Objective

• Evaluate the performance of two green roof models, EcoRoof and WRF-MLGR, against observations from a monitored green roof in London, Ontario.
• Assess the ability of each model to reproduce fluxes (Qe, Qg) and Tsurf during multi-month periods.
• Investigate how model performance changes across different moisture regimes and radiative forcing conditions.

Study Configuration

• Spatial Scale: Local scale, focusing on a single green roof in London, Ontario, Canada.
• Temporal Scale: Multi-month periods, with two analysis periods of 62 days each (July-August and September-October).

Methodology and Data

• Models used:
  • EcoRoof (ER): a stand-alone version of the original model found within EnergyPlus v8.6.0.
  • WRF-MLGR: a multilayer green roof parametrization embedded in WRF v4.3.1.
• Data sources: Observations from a monitored green roof in London, Ontario, including lysimeter-derived Qe and Qg, Tsurf, and substrate water content (SWC).

Main Results

• EcoRoof (ER) shows similar performance across both periods, with mean RMSE values of approximately 56-60 Wm^-2 for Qe, 3-4°C for Tsurf, 22 Wm^-2 for Qg, and 0.03 m^3 m^-3 for SWC.
• WRF-MLGR exhibits larger errors for Qg (mean RMSE exceeding 100 Wm^-2) and SWC (0.06 m^3 m^-3), along with higher Tsurf deviations (~5°C).
• ER provides a more consistent representation of the daytime energy balance, while WRF-MLGR shows structural biases in surface heating.

Contributions

• This study highlights the importance of process-level evaluation for urban climate applications and underscores the need for continued model development to address structural limitations in green roof parametrizations.
• The results provide insights into the performance of two widely used green roof models, EcoRoof and WRF-MLGR, against observations from a monitored green roof.

Funding

• This research was funded by [insert funding agency or project name].

Citation

@article{Mendoza2026Processbased,
  author = {Mendoza, Maria Martinez and Saeedi, Alireza and Voogt, James and Krayenhoff, E. Scott},
  title = {Process-based evaluation of green roof models for assessment of heat mitigation efficacy in WRF (v4.3.1) and EnergyPlus (v8.6.0)},
  journal = {Geoscientific model development},
  year = {2026},
  doi = {10.5194/gmd-19-9729-2026},
  url = {https://doi.org/10.5194/gmd-19-9729-2026}
}