• Aliabadi, A. A., M. Moradi, D. Clement, W. D. Lubitz, and B. Gharabaghi, 2019: Flow and temperature dynamics in an urban canyon under a comprehensive set of wind directions, wind speeds, and thermal stability conditions. Environ. Fluid Mech., 19, 81109, https://doi.org/10.1007/s10652-018-9606-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Aligo, E. A., W. A. Gallus Jr., and M. Segal, 2009: On the impact of WRF Model vertical grid resolution on Midwest summer rainfall forecasts. Wea. Forecasting, 24, 575594, https://doi.org/10.1175/2008WAF2007101.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Awan, N. K., H. Truhetz, and A. Gobiet, 2011: Parameterization-induced error characteristics of MM5 and WRF operated in climate mode over the Alpine region: An ensemble-based analysis. J. Climate, 24, 31073123, https://doi.org/10.1175/2011JCLI3674.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Byerlay, R. A. E., M. K. Nambiar, A. Nazem, M. R. Nahian, M. Biglarbegian, and A. A. Aliabadi, 2020: Measurement of land surface temperature from oblique angle airborne thermal camera observations. Int. J. Remote Sens., 41, 31193146, https://doi.org/10.1080/01431161.2019.1699672.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Carvalho, D., A. Rocha, M. Gómez-Gesteira, and C. Santos, 2012: A sensitivity study of the WRF model in wind simulation for an area of high wind energy. Environ. Modell. Software, 33, 2334, https://doi.org/10.1016/j.envsoft.2012.01.019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cerlini, P. B., K. A. Emanuel, and E. Todini, 2005: Orographic effects on convective precipitation and space-time rainfall variability: Preliminary results. Hydrol. Earth Syst. Sci., 9, 285299, https://doi.org/10.5194/hess-9-285-2005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cheng, F.-Y., Y.-C. Hsu, P.-L. Lin, and T.-H. Lin, 2013: Investigation of the effects of different land use and land cover patterns on mesoscale meteorological simulations in the Taiwan area. J. Appl. Meteor. Climatol., 52, 570587, https://doi.org/10.1175/JAMC-D-12-0109.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cheng, W. Y., and W. J. Steenburgh, 2005: Evaluation of surface sensible weather forecasts by the WRF and the eta models over the western United States. Wea. Forecasting, 20, 812821, https://doi.org/10.1175/WAF885.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chu, X., L. Xue, B. Geerts, R. Rasmussen, and D. Breed, 2014: A case study of radar observations and WRF LES simulations of the impact of ground-based glaciogenic seeding on orographic clouds and precipitation. Part I: Observations and model validations. J. Appl. Meteor. Climatol., 53, 22642286, https://doi.org/10.1175/JAMC-D-14-0017.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Clements, C. B., C. D. Whiteman, and J. D. Horel, 2003: Cold-air-pool structure and evolution in a mountain basin: Peter Sinks, Utah. J. Appl. Meteor., 42, 752768, https://doi.org/10.1175/1520-0450(2003)042<0752:CSAEIA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • De Meij, A., and J. F. Vinuesa, 2014: Impact of SRTM and Corine land cover data on meteorological parameters using WRF. Atmos. Res., 143, 351370, https://doi.org/10.1016/j.atmosres.2014.03.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Diagne, M., M. David, J. Boland, N. Schmutz, and P. Lauret, 2014: Post-processing of solar irradiance forecasts from WRF model at Reunion Island. Sol. Energy, 105, 99108, https://doi.org/10.1016/j.solener.2014.03.016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dudhia, J., 1989: Numerical study of convection observed during the Winter Monsoon Experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46, 30773107, https://doi.org/10.1175/1520-0469(1989)046<3077:NSOCOD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fernández-González, S., M. L. Martín, E. García-Ortega, A. Merino, J. Lorenzana, J. L. Sánchez, F. Valero, and J. S. Rodrigo, 2018: Sensitivity analysis of the WRF Model: Wind-resource assessment for complex terrain. J. Appl. Meteor. Climatol., 57, 733753, https://doi.org/10.1175/JAMC-D-17-0121.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flesch, T. K., J. D. Wilson, and E. Yee, 1995: Backward-time Lagrangian stochastic dispersion models and their application to estimate gaseous emissions. J. Appl. Meteor., 34, 13201332, https://doi.org/10.1175/1520-0450(1995)034<1320:BTLSDM>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flesch, T. K., J. D. Wilson, L. A. Harper, and B. P. Crenna, 2005: Estimating gas emissions from a farm with an inverse-dispersion technique. Atmos. Environ., 39, 48634874, https://doi.org/10.1016/j.atmosenv.2005.04.032.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flesch, T. K., J. D. Wilson, L. A. Harper, R. W. Todd, and N. A. Cole, 2007: Determining ammonia emissions from a cattle feedlot with an inverse dispersion technique. Agric. For. Meteor., 144, 139155, https://doi.org/10.1016/j.agrformet.2007.02.006.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Friedl, M. A., D. Sulla-Menashe, B. Tan, A. Schneider, N. Ramankutty, A. Sibley, and X. Huang, 2010: MODIS Collection 5 global land cover: Algorithm refinements and characterization of new datasets. Remote Sens. Environ., 114, 168182, https://doi.org/10.1016/j.rse.2009.08.016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gilliam, R. C., and J. E. Pleim, 2010: Performance assessment of new land surface and planetary boundary layer physics in the WRF-ARW. J. Appl. Meteor. Climatol., 49, 760774, https://doi.org/10.1175/2009JAMC2126.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gu, H., J. Jin, Y. Wu, M. B. Ek, and Z. M. Subin, 2015: Calibration and validation of lake surface temperature simulations with the coupled WRF-lake model. Climatic Change, 129, 471483, https://doi.org/10.1007/s10584-013-0978-y.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hari Prasad, K. B. R. R., C. V. Srinivas, T. N. Rao, C. V. Naidu, and R. Baskaran, 2017: Performance of WRF in simulating terrain induced flows and atmospheric boundary layer characteristics over the tropical station Gadanki. Atmos. Res., 185, 101117, https://doi.org/10.1016/j.atmosres.2016.10.020.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • He, J. J., Y. Yu, L. J. Yu, N. Liu, and S. P. Zhao, 2017: Impacts of uncertainty in land surface information on simulated surface temperature and precipitation over China. Int. J. Climatol., 37, 829847, https://doi.org/10.1002/joc.5041.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., 2012: Orographic effects on precipitating clouds. Rev. Geophys., 50, RG1001, https://doi.org/10.1029/2011RG000365.

  • Hu, N., T. K. Flesch, J. D. Wilson, V. S. Baron, and J. A. Basarab, 2016: Refining an inverse dispersion method to quantify gas sources on rolling terrain. Agric. For. Meteor., 225, 17, https://doi.org/10.1016/j.agrformet.2016.05.007.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hu, X.-M., J. W. Nielsen-Gammon, and F. Zhang, 2010: Evaluation of three planetary boundary layer schemes in the WRF Model. J. Appl. Meteor. Climatol., 49, 18311844, https://doi.org/10.1175/2010JAMC2432.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jiménez, P. A., and J. Dudhia, 2012: Improving the representation of resolved and unresolved topographic effects on surface wind in the WRF Model. J. Appl. Meteor. Climatol., 51, 300316, https://doi.org/10.1175/JAMC-D-11-084.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lehner, M., and Coauthors, 2016: The METCRAX II field experiment: A study of downslope windstorm-type flows in Arizona’s Meteor Crater. Bull. Amer. Meteor. Soc., 97, 217235, https://doi.org/10.1175/BAMS-D-14-00238.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liggio, J., and Coauthors, 2019: Measured Canadian oil sands CO2 emissions are higher than estimates made using internationally recommended methods. Nat. Commun., 10, 1863, https://doi.org/10.1038/s41467-019-09714-9.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lin, Y.-L., R. D. Farley, and H. D. Orville, 1983: Bulk parameterization of the snow field in a cloud model. J. Climate Appl. Meteor., 22, 10651092, https://doi.org/10.1175/1520-0450(1983)022<1065:BPOTSF>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lorente-Plazas, R., P. A. Jiménez, J. Dudhia, and J. P. Montávez, 2016: Evaluating and improving the impact of the atmospheric stability and orography on surface winds in the WRF Model. Mon. Wea. Rev., 144, 26852693, https://doi.org/10.1175/MWR-D-15-0449.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mass, C. F., D. Ovens, K. Westrick, and B. A. Colle, 2002: Does increasing horizontal resolution produce more skillful forecasts? The results of two years of real-time numerical weather prediction over the Pacific Northwest. Bull. Amer. Meteor. Soc., 83, 407430, https://doi.org/10.1175/1520-0477(2002)083<0407:DIHRPM>2.3.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mirocha, J. D., M. D. Simpson, J. D. Fast, L. K. Berg, and R. Baskett, 2016: Investigation of boundary-layer wind predictions during nocturnal low-level jet events using the Weather Research and Forecasting model. Wind Energy, 19, 739762, https://doi.org/10.1002/we.1862.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mughal, M. O., M. Lynch, F. Yu, B. McGann, F. Jeanneret, and J. Sutton, 2017: Wind modelling, validation and sensitivity study using Weather Research and Forecasting model in complex terrain. Environ. Modell. Software, 90, 107125, https://doi.org/10.1016/j.envsoft.2017.01.009.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nossent, J., P. Elsen, and W. Bauwens, 2011: Sobol’ sensitivity analysis of a complex environmental model. Environ. Modell. Software, 26, 15151525, https://doi.org/10.1016/j.envsoft.2011.08.010.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pontoppidan, M., J. Reuder, S. Mayer, and E. W. Kolstad, 2017: Downscaling an intense precipitation event in complex terrain: The importance of high grid resolution. Tellus, 69A, 1271561, https://doi.org/10.1080/16000870.2016.1271561.

    • Search Google Scholar
    • Export Citation
  • Powers, J. G., and Coauthors, 2017: The Weather Research and Forecasting Model: Overview, system efforts, and future directions. Bull. Amer. Meteor. Soc., 98, 17171737, https://doi.org/10.1175/BAMS-D-15-00308.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Qiu, X., I. Cheng, F. Yang, E. Horb, L. Zhang, and T. Harner, 2018: Emissions databases for polycyclic aromatic compounds in the Canadian Athabasca oil sands region—Development using current knowledge and evaluation with passive sampling and air dispersion modelling data. Atmos. Chem. Phys., 18, 34573467, https://doi.org/10.5194/acp-18-3457-2018.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Saide, P. E., G. R. Carmichael, S. N. Spak, L. Gallardo, A. E. Osses, M. A. Mena-Carrasco, and M. Pagowski, 2011: Forecasting urban PM10 and PM2.5 pollution episodes in very stable nocturnal conditions and complex terrain using WRF-Chem CO tracer model. Atmos. Environ., 45, 27692780, https://doi.org/10.1016/j.atmosenv.2011.02.001.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Storm, B., J. Dudhia, S. Basu, A. Swift, and I. Giammanco, 2009: Evaluation of the Weather Research and Forecasting Model on forecasting low-level jets: Implications for wind energy. Wind Energy, 12, 8190, https://doi.org/10.1002/we.288.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Subin, Z. M., W. J. Riley, and D. Mironov, 2012: An improved lake model for climate simulations: Model structure, evaluation, and sensitivity analyses in CESM1. J. Adv. Model. Earth Syst., 4, M02001, https://doi.org/10.1029/2011MS000072.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Talbot, C., E. Bou-Zeid, and J. Smith, 2012: Nested mesoscale large-eddy simulations with WRF: Performance in real test cases. J. Hydrometeor., 13, 14211441, https://doi.org/10.1175/JHM-D-11-048.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Taylor, C. M., E. F. Lambin, N. Stephenne, R. J. Harding, and R. L. Essery, 2002: The influence of land use change on climate in the Sahel. J. Climate, 15, 36153629, https://doi.org/10.1175/1520-0442(2002)015<3615:TIOLUC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Walpole, R. E., R. E. Myers, S. L. Myers, and K. Ye, 2002: Probability & Statistics for Engineers & Scientists. 7th ed. Prentice Hall, 730 pp.

    • Search Google Scholar
    • Export Citation
  • Whiteman, C. D., T. Haiden, B. Pospichal, S. Eisenbach, and R. Steinacker, 2004: Minimum temperatures, diurnal temperature ranges, and temperature inversions in limestone sinkholes of different sizes and shapes. J. Appl. Meteor., 43, 12241236, https://doi.org/10.1175/1520-0450(2004)043<1224:MTDTRA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Whiteman, C. D., and Coauthors, 2008: METCRAX 2006: Meteorological experiments in Arizona’s Meteor Crater. Bull. Amer. Meteor. Soc., 89, 16651680, https://doi.org/10.1175/2008BAMS2574.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, L., H. Liu, Q. Du, and L. Wang, 2016: Evaluation of the WRF-lake model over a highland freshwater lake in southwest China. J. Geophys. Res. Atmos., 121, 13 98914 005, https://doi.org/10.1002/2016JD025396.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yan, H., and W. A. Gallus Jr., 2016: An evaluation of QPF from the WRF, NAM, and GFS models using multiple verification methods over a small domain. Wea. Forecasting, 31, 13631379, https://doi.org/10.1175/WAF-D-16-0020.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yáñez-Morroni, G., J. Gironás, M. Caneo, R. Delgado, and R. Garreaud, 2018: Using the Weather Research and Forecasting (WRF) model for precipitation forecasting in an Andean region with complex topography. Atmosphere, 9, 304, https://doi.org/10.3390/atmos9080304.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, C., H. Lin, M. Chen, and L. Yang, 2014: Scale matching of multiscale digital elevation model (DEM) data and the Weather Research and Forecasting (WRF) model: A case study of meteorological simulation in Hong Kong. Arab. J. Geosci., 7, 22152223, https://doi.org/10.1007/s12517-014-1273-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, D.-L., and W.-Z. Zheng, 2004: Diurnal cycles of surface winds and temperatures as simulated by five boundary layer parameterizations. J. Appl. Meteor., 43, 157169, https://doi.org/10.1175/1520-0450(2004)043<0157:DCOSWA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zheng, Y., K. Alapaty, J. A. Herwehe, A. D. Del Genio, and D. Niyogi, 2016: Improving high-resolution weather forecasts using the Weather Research and Forecasting (WRF) Model with an updated Kain–Fritsch scheme. Mon. Wea. Rev., 144, 833860, https://doi.org/10.1175/MWR-D-15-0005.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhu, R., and D. Xu, 2004: Multi-scale turbulent planetary boundary layer parameterization in mesoscale numerical simulation. Quart. J. Appl. Meteor., 15, 543555.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a) Map of the regional area where the mining facility is located; (b) location of the meteorological instruments deployed at the site for model comparison, color coded with surface height above sea level.

  • View in gallery

    WRF simulation domains centered at the mining facility, color coded with surface height above sea level; the largest domain (domain 1) covers parts of Alberta, British Columbia, Saskatchewan, and Manitoba provinces. It also covers parts of Yukon, Northwest, and Nunavut Territories. The smallest domain (domain 5) contains primarily the mining facility. The main population centers are marked on the figure.

  • View in gallery

    Surface height from sea level: GTOPO30 is the nonmodified land; SRTM 1 s with lidar contains land modifications showing the mine and the tailings pond.

  • View in gallery

    The two land-use configurations for WRF simulations; mine (barren = 16), pond (lake =21), plant and lodge (urban and built-up = 13), and other (grassland = 10); lakes are considered as class 21 during simulation but are not color coded.

  • View in gallery

    Selected vertical profiles of (a) mean wind speed and (b) mean potential temperature measured by the TANAB in the mine (on 18 and 24 May 2018) and on the east side of the pond (on 30 May 2018), grouped in 2-h time intervals; the vertical axis shows height above ground level; each mean was calculated over a 5-min sampling period.

  • View in gallery

    Effects of topography and land-use changes on wind speed near the surface; the instruments were set up at 10 m except for the TANAB; the average value for the three dates at every hour is shown for both the model and the observations.

  • View in gallery

    Effects of topography and land-use changes on 10-m wind vector magnitude and direction at 0200 and 1400 MST 18 May 2018; black lines show mine and pond perimeters.

  • View in gallery

    Effects of topography and land-use changes on the absolute error of wind direction at 10 m; the average value of the absolute error of wind direction for the three dates at every hour is shown.

  • View in gallery

    Effects of topography and land-use changes on temperature at 2 m; the average value for the three dates at every hour is shown for both the model and the observations.

  • View in gallery

    Effects of topography and land-use changes on 2-m potential temperature at 0200 and 1400 MST 18 May 2018; black lines show mine and pond perimeters.

  • View in gallery

    Effects of topography and land-use changes on turbulent sensible heat flux at 10 m; the average value for the three dates at every hour is shown for both the model and the observations.

  • View in gallery

    Effects of topography and land-use changes on relative humidity at 2 m; the average value for the three dates at every hour is shown for both the model and the observations.

All Time Past Year Past 30 Days
Abstract Views 120 0 0
Full Text Views 404 260 14
PDF Downloads 386 247 13

Complex Meteorology over a Complex Mining Facility: Assessment of Topography, Land Use, and Grid Spacing Modifications in WRF

View More View Less
  • 1 School of Engineering, University of Guelph, Guelph, Ontario, Canada
  • | 2 RWDI, Calgary, Alberta, Canada
  • | 3 Southern Alberta Institute of Technology, Calgary, Alberta, Canada
  • | 4 School of Engineering, University of Guelph, Guelph, Ontario, Canada
© Get Permissions
Free access

Abstract

The performance of the Weather Research and Forecasting (WRF) Model is evaluated in predicting the meteorological conditions over a complex open-pit mining facility in northern Canada in support of more accurate operational reporting of area-fugitive greenhouse gas emission fluxes from such facilities. WRF is studied in a series of sensitivity tests by varying topography, land use, and horizontal and vertical grid spacings to arrive at optimum configurations for reducing modeling biases in comparison with field meteorological observations. Overall, WRF shows a better performance when accounting for the mine topography and modified land use. As a result, the model biases reduce from 1.10 to 0.08 m s−1, from 1.04 to 0.50 m s−1, from 0.98 to 0.32 K, and from 45.7 to 17.3 W m−2, for near-surface wind speed, boundary layer wind speed, near-surface potential temperature, and turbulent sensible heat flux, respectively. Refining the model horizontal and vertical grid spacings results in bias reductions from 3.31 to 0.08 and from 0.80 to −0.11 m s−1 for near-surface and boundary layer wind speeds, respectively. The simulation results also agree with previous observations of meteorological effects on enclosed Earth depressions, characterized by formation of a cool pool of air, reduced wind speeds, and horizontal wind circulations at the bottom of the depression under thermally stable conditions. The results suggest that such configurations for WRF are necessary to arrive at more accurate meteorological predictions over complex open-pit mining terrains with similar features.

© 2020 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Amir A. Aliabadi, aliabadi@uoguelph.ca

Abstract

The performance of the Weather Research and Forecasting (WRF) Model is evaluated in predicting the meteorological conditions over a complex open-pit mining facility in northern Canada in support of more accurate operational reporting of area-fugitive greenhouse gas emission fluxes from such facilities. WRF is studied in a series of sensitivity tests by varying topography, land use, and horizontal and vertical grid spacings to arrive at optimum configurations for reducing modeling biases in comparison with field meteorological observations. Overall, WRF shows a better performance when accounting for the mine topography and modified land use. As a result, the model biases reduce from 1.10 to 0.08 m s−1, from 1.04 to 0.50 m s−1, from 0.98 to 0.32 K, and from 45.7 to 17.3 W m−2, for near-surface wind speed, boundary layer wind speed, near-surface potential temperature, and turbulent sensible heat flux, respectively. Refining the model horizontal and vertical grid spacings results in bias reductions from 3.31 to 0.08 and from 0.80 to −0.11 m s−1 for near-surface and boundary layer wind speeds, respectively. The simulation results also agree with previous observations of meteorological effects on enclosed Earth depressions, characterized by formation of a cool pool of air, reduced wind speeds, and horizontal wind circulations at the bottom of the depression under thermally stable conditions. The results suggest that such configurations for WRF are necessary to arrive at more accurate meteorological predictions over complex open-pit mining terrains with similar features.

© 2020 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Amir A. Aliabadi, aliabadi@uoguelph.ca

1. Introduction

Quantifying area-fugitive emissions from complex terrains, such as an open-pit mine, poses a challenge to operational techniques used in emission reporting. For example, a study by Liggio et al. (2019) showed that area-fugitive carbon dioxide emission intensities from open-pit mines measured in situ by an aircraft deviate by 13%–123% from those estimated from emission inventory datasets. Because of the scale of open-pit mining operations, accurate observations of area-fugitive emissions are impractical, so recent effort has been made to develop operational modeling tools to estimate emissions. For instance, inverse dispersion modeling (IDM) employs backward Lagrangian calculations to estimate an emission source strength distributed over an area given limited point observations downwind. Although computationally fast (Flesch et al. 1995), the IDM method suffers in complex topographies and land-use environments where horizontal homogeneity in meteorological fields cannot be assumed (Flesch et al. 2005, 2007; Hu et al. 2016). Recent availability of computational power has enabled estimation of area-fugitive emissions by forward dispersion modeling, where a spatiotemporally resolved input source is provided to the model to predict the emission flux downwind. For example, the California Puff Model (CALPUFF) is a multilayer, multispecies, and non-steady-state puff dispersion model, which simulates the effects of time- and space-varying meteorological conditions on pollution transport. Operational CALPUFF is usually run with a diagnostic three-dimensional meteorological model, whereas ingestion of meteorological fields from mesoscale models, such as the Weather Research and Forecasting (WRF) Model, can greatly enhance CALPUFF’s performance for future operational use (Qiu et al. 2018).

The natural topographical examples of Earth similar to the depression of an open-pit mine are the Arizona’s Meteor Crater (Whiteman et al. 2008; Lehner et al. 2016), Peter Sinks in Utah (Clements et al. 2003), and the Gruenloch doline in Austria (Whiteman et al. 2004), all of which have been studied in detail. Meteorological features of such depressions are different from flat terrain and valleys. For instance, a temperature-stratified, calm, and cool pool of air develops at the bottom of such depressions during the night (Clements et al. 2003; Whiteman et al. 2004, 2008; Lehner et al. 2016). In comparison to flat terrain and valleys, the meteorological conditions inside the depression exhibit reduced slope flows, reduced advective transfer with outside of the depression, reduced turbulent sensible heat flux at the bottom, change of wind direction from the cool pool to elevations outside the depression, and formation of weak and intermittent turbulent jets on the depression walls near the ground (Clements et al. 2003; Whiteman et al. 2004). In addition, standing waves or oscillating temperature fields have been observed in such depressions, and the temperature stratification is observed to break under high wind conditions (Whiteman et al. 2008; Lehner et al. 2016). The meteorological conditions of such depressions are understood to be influenced by synoptic events, the seasonal weather variation, topography, and radiative heat transfer between the depression and the sky, which is a function of depression aspect ratio (Clements et al. 2003; Whiteman et al. 2004). It is yet to be investigated if open-pit mines feature similar meteorological conditions, particularly given the confounding influence of nearby industrial operations and water bodies such as tailings ponds.

The success of the WRF Model with applications in complex topography and land use in the open-pit mining facilities is yet to be investigated in more detail. Mesoscale models require land surface, surface layer, and planetary boundary layer (PBL) parameterizations to represent the transfer of heat, moisture, and momentum between the surface and atmosphere (Gilliam and Pleim 2010). Historically, Cheng and Steenburgh (2005) have reported that WRF exhibits a positive wind speed bias over land at the lowest grid layer, while recently Fernández-González et al. (2018) reported a negative bias over land in complex terrains at the lowest grid layer possibly due to smoothened orography in mesoscale models that enhance wind gradients in the lower levels of the PBL. Such biases still exist in the latest versions that represent a limitation for the high demand of accurate wind estimations by different sectors. Since current atmospheric models present an extensive spectrum of configuration options and parameters, selecting the optimum configuration among these alternatives has its own inherent challenges (Nossent et al. 2011). One of the major limitations of the WRF Model at very fine grid spacing is the inaccuracy of the real terrain representation (Jiménez and Dudhia 2012). For accurate predictions in meteorology, topographic and land-use data are important inputs for weather forecasting models, and high-resolution terrain data are required for the expression of complex characteristics of various regions (Lorente-Plazas et al. 2016). The complex terrains influence the meteorological conditions by altering various meteorological variables such as wind speed and direction, albedo, and surface heat flux (Cerlini et al. 2005). It has been found that increasing topographic and land-use resolution of input datasets, along with using finer grid spacing, is more important than just refining the model grid spacing toward more accurate simulations (Zhu and Xu 2004). The regional circulations are affected by topographic properties that play an important role in a numerical model in altering the rainfall, airflow, and temperature distribution (Houze 2012; Yáñez-Morroni et al. 2018). The land-use data play a significant role in altering the local and regional weather by affecting the aerodynamic and thermodynamic interactions between the surface and atmosphere (Taylor et al. 2002). The wind speed and the temperature can change due to the impacts of high frictional forces and high aerodynamic roughness length, when higher topographic and land-use resolution data are applied (De Meij and Vinuesa 2014). The temperature and relative humidity show better agreement with experimental observations when high-resolution topographic data are used (Zhang et al. 2014).

Sensitivity tests using WRF can result in improved performance for simulations of meteorological conditions over complex terrains. This entails selecting various schemes, different grid configurations, alternative forcing datasets, and assimilating in situ observations. The error minimization in model variables can be achieved by testing and choosing a suitable numerical and physical configuration for the region of interest. Although the fine horizontal and vertical grid spacings may lead to better reproduction of finescale meteorological processes, this may not necessarily be true due to uncertainties in the overall performance of the various physical parameterizations and their responses to grid spacing (Carvalho et al. 2012). However, the sensitivity studies of Mass et al. (2002) and Pontoppidan et al. (2017) showed that refining the grid spacing improved the model simulations in complex orographic zones. Powers et al. (2017) described several physical parameterization schemes available for microphysics, radiation, and clouds as well as boundary layer schemes including the surface layer, the PBL, and the land surface model. Such schemes have nonlinear interactions with each other and with the dynamical core of the model. Therefore, it becomes challenging to optimize the model due to these complex relationships. Besides physical parameterization schemes and unconfined empirical parameters within these schemes, there are other sources of errors in the numerical model. Such model errors arise because of the interdependence of different numerical solvers, domain sizes, regional topography, boundary conditions, terrain, vegetation characteristics, and grid configuration in both horizontal and vertical directions (Awan et al. 2011).

Although several studies have been conducted to evaluate the performance of the WRF Model by using high-resolution topography and land-use datasets, there has been a lack of research involving the use of WRF for meteorological predictions over open-pit mining facilities. In particular, there is a need to better understand WRF, its uncertainties, and its potential for more accurate and operational quantification of area-fugitive emissions from open-pit mining facilities. This is the central focus of this paper. This paper takes the first step to investigate the meteorological conditions over an open-pit mining facility, and future work will focus on dispersion of greenhouse gas emissions from the facility.

The effects of high topographic resolution, modified land use, and grid configuration in the horizontal and vertical directions, on the accuracy of the model results require more research. Moreover, most of the experimental observations are conducted near-surface by using various meteorological instruments on a meteorological tower. Hence, the main objectives of this study are to observe the meteorological variables (both near-surface and aloft) over a complex mining facility located in boreal forests in northern Canada, and to perform a comparative analysis between the model and the observations. Further, the model is investigated in a series of sensitivity studies to assess the effects of terrain topographic resolution, land-use change, and grid configuration, on the accuracy of the model.

The structure of the paper is as follows. Section 2 provides the methodology of the study. The details of the field meteorological campaign and the experimental observations are outlined in section 2a. Section 2b supplies the details for the WRF Model including domain configurations, initialization, topography, land use, grid spacing, PBL parameterization, and evaluation. Section 3 gives the results and discussion. Effects of topography and land use on the model outputs are discussed in section 3a. The effects of grid spacing are discussed in section 3b. Section 4 furnishes the summary and conclusions.

2. Methods

a. Experimental observations

The experimental measurements were performed in an open-pit mining facility in northern Canada in May 2018. The facility is located near the Wood Buffalo National Park of Canada (Fig. 1a). As shown in Fig. 1b, the facility includes a tailings pond, which is an area of refused mining waste where the waterborne refuse material is pumped, and most of the outlets of the pumps are located near the barren area on the east side of the pond. Open-pit mining excavations are primarily conducted over the mine area. The mine is approximately 100 m deep, with a width-to-depth aspect ratio of greater than 20.

Fig. 1.
Fig. 1.

(a) Map of the regional area where the mining facility is located; (b) location of the meteorological instruments deployed at the site for model comparison, color coded with surface height above sea level.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

Figure 1b also outlines the locations of the instruments deployed. Because of model computational cost and based on the availability of experimental datasets, no more than three days of simulations were possible. Three days, 18, 24, and 30 May 2018, were chosen for observation-model comparison. These days correspond to clear sky conditions without synoptic events. The following instruments were used to determine wind speed, temperature, relative humidity, and turbulent sensible heat flux at heights at or below 200 m above ground level (i.e., near surface) or at heights above 200 m above ground level within the PBL in the mining facility.

1) Sodar

A PA-5 sonic detection and ranging (sodar) instrument by Remtech, Inc. (http://www.remtechinc.com), was used to measure wind speed and direction from 100 to 2700 m with an output frequency of 60 min. This acoustic wind profiler has the capacity of measuring wind speed from 0 to 40 m s−1 with accuracies of ±0.2 m s−1 and ±3° for wind speeds over 6 and 2 m s−1, respectively. It was located to the west side of a tailings pond.

2) TANAB

The customized Tethered and Navigated Air Blimp (TANAB) contains a microclimate sensor called TriSonica Mini weather station by Applied Technologies, Inc. (http://www.apptech.com), to measure wind speed, wind direction, pressure, temperature, and relative humidity with a sampling frequency of 10 Hz (Byerlay et al. 2020). It is capable of measuring wind speed from 0 to 30 m s−1, temperature from 248.15 to 353.15 K, pressure from 50 to 115 kPa, and relative humidity from 0% to 100%. The measurement resolution of this miniature weather station is ±0.1 m s−1 for wind speed, ±1° for wind direction, and ±0.1 K for temperature. Moreover, the accuracy of measurement is ±0.1 m s−1 for wind speed, ±1° for wind direction, and ±2 K for temperature. The weather station was rigorously calibrated in a wind tunnel. The TANAB was launched up to an altitude of 200 m from the surface both in the mine (on 18 and 24 May 2018) and near the east side of the pond (on 30 May 2018). Each launch took about 1 h but was repeated each day. Meteorological variables were statistically sampled every 5 min to produce means of wind speed and potential temperature as a function of time of day and height.

3) 3D ultrasonic anemometers

Two CSAT 3B ultrasonic anemometers from Campbell Scientific, Inc. (https://www.campbellsci.ca), were used to measure the 3D wind components and temperature at a sampling frequency of 10 Hz as recommended by Aliabadi et al. (2019). The anemometer has the capability of measuring wind speed up to 30 m s−1 and temperature from 243.15 to 323.15 K. The measurement resolutions of these ultrasonic anemometers are ±0.001 m s−1 for horizontal wind, ±0.0005 m s−1 for vertical wind, ±0.058° for wind direction, and ±0.002 K for temperature. Moreover, the accuracies are ±0.08 m s−1 for horizontal wind, ±0.04 m s−1 for vertical wind, and ±10° for wind direction. These ultrasonic anemometers were located to the north side of the mine (3D Ultrasonic 1) and southwest of the pond (3D Ultrasonic 2) at 10-m elevation on meteorological towers.

4) 2D ultrasonic anemometers

Three model 86004 2D ultrasonic anemometers from R. M. Young Company (http://www.youngusa.com) were operated to measure wind speed and wind direction at a sampling frequency of 10 Hz. These anemometers can measure wind speed in a range from 0 to 65 m s−1 with an accuracy of ±0.2 m s−1 and wind direction with an accuracy of ±2°. The measurement resolution of wind speed and wind direction are ±0.01 m s−1 and ±0.1°, respectively. These 2D ultrasonic anemometers were located to the southwest of the mine (2D Ultrasonic 1) and east (2D Ultrasonic 2) and northwest of the pond (2D Ultrasonic 3). They were installed at 10-m elevation on meteorological towers.

5) Weather stations

Two weather stations were also operational. For the first weather station, a Gill 3-cup anemometer and model 41382 relative humidity/temperature probe from R. M. Young Company were set up to measure wind speed, wind direction, temperature, and relative humidity at 1 Hz. The temperature and relative humidity probe can measure temperature from 223.15 to 323.15 K with an accuracy of ±0.3 K and relative humidity from 0% to 100% with an accuracy of ±1%. The cup anemometer and the relative humidity/temperature probe were located at the north side of the mine on a trailer that was 2 m above the ground. The second weather station was located at the southeast corner of the mining facility. The meteorological data from this station were downloaded from the Wood Buffalo Environmental Association (https://wbea.org/).

6) Lidar topographical measurements

Light detection and ranging (lidar) observations were performed by Clean Harbors Company (http://www.airborneimaginginc.com) to determine the actual topography of the mining facility. Airborne imaging was performed using a fixed-wing drone. Any error in the altitude of the aircraft (roll, pitch, and heading) was observed and corrected for within system specifications. To statistically quantify the accuracy, the lidar elevations were compared with independently surveyed ground points. This lidar observation has an accuracy of ±0.3 m in the horizontal direction and ±0.1 m in the vertical direction.

b. WRF Model

A particular distribution of WRF is used in this study, titled Unified Environmental Modeling System (UEMS), version 18.1.1 (http://strc.comet.ucar.edu/software/uems/). The UEMS allows for the acquisition of multiple initialization datasets via the Nonhydrostatic Forecast System and hypertext transfer protocol. For this study, each WRF simulation was run on 25 CPUs. Simulation times varied from 6 h to 14 days for each case depending on the horizontal and vertical grid spacings.

1) Domain configurations

WRF has been tested with various configurations including the way of nesting, grid spacing, time-invariant data resolution, and geographical shifts in the domain center. The minimum grid distance of the nested domains should be at least four grid cells from the parent boundary. Figure 2 and Table 1 illustrate the size and configuration of the domains, respectively, where domain 1 is the largest domain and domain 5 is the smallest one (Fig. 1b shows domain 5). The domains consist of five nested levels. The largest domain (domain 1) covers most of northern and western Canada, and the smallest domain (domain 5) contains primarily the mining facility.

Fig. 2.
Fig. 2.

WRF simulation domains centered at the mining facility, color coded with surface height above sea level; the largest domain (domain 1) covers parts of Alberta, British Columbia, Saskatchewan, and Manitoba provinces. It also covers parts of Yukon, Northwest, and Nunavut Territories. The smallest domain (domain 5) contains primarily the mining facility. The main population centers are marked on the figure.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

Table 1.

Domain configurations and associated parameters.

Table 1.

2) Initialization

The WRF simulations are executed for three different days on 18, 24, and 30 May 2018, for 36 h in each day including a spinup time of 12 h. Different configurations have been studied in sensitivity tests varying the topographic resolution, land use, horizontal grid spacing, and number of vertical levels. The simulations are run with the nonhydrostatic Advanced Research WRF (ARW) solver, with the one-way nesting method being chosen for better stability (Saide et al. 2011). The Global Forecasting System (https://www.ncdc.noaa.gov/data-access/model-data/model-datasets/global-forcast-system-gfs) dataset has been used for the simulations to provide initial and time-varying boundary conditions to WRF at 3-h time resolution and 0.5° spatial grid spacing (Diagne et al. 2014; Yan and Gallus 2016; Powers et al. 2017). The other implemented physics are the Lin scheme for microphysics, multiscale Kain–Fritsch scheme for cumulus physics, and Dudhia scheme for longwave and shortwave radiation (Lin et al. 1983; Dudhia 1989; Zheng et al. 2016). A ratio of ⅓ for both time step and horizontal grid spacing in successive nested grids is used as suggested by Powers et al. (2017). For example, at fine horizontal and midscale vertical grid spacings, the model time steps were on the order of 200–2 s for domains 1–5, respectively.

3) Topography and land use

As shown in Fig. 3, the default Global 30-arc-s elevation dataset (GTOPO30) and the Shuttle Radar Topography Mission 1-arc-s (SRTM 1 s) datasets are used to modify the resolution of topography in domains 4 and 5. GTOPO30 is a global dataset covering the full extent of Earth. As a low-resolution topographic initialization, the GTOPO30 dataset has a topographic horizontal resolution of 900 m × 900 m. On the other hand, the SRTM 1-s dataset contains digital elevation models on a near-global scale to generate the most complete high-resolution digital topographic database of Earth. The SRTM 1-s dataset has a topographic resolution of 30 m × 30 m (De Meij and Vinuesa 2014). Moreover, in this study, the SRTM 1-s dataset is corrected with the lidar topographical measurements to include the actual topography of the mining facility in WRF simulations.

Fig. 3.
Fig. 3.

Surface height from sea level: GTOPO30 is the nonmodified land; SRTM 1 s with lidar contains land modifications showing the mine and the tailings pond.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

Land-use (LU) configurations in WRF describe surface variables, such as roughness, permeability, and so on, that are important drivers for atmospheric dynamics. In the mining facility, changes in land surface must be correctly accounted for to be able to realistically simulate atmospheric transport and dynamics. It has been demonstrated that LU data can have a significant effect on meteorological simulations (Cheng et al. 2013; Xu et al. 2016). Two LU configurations are considered for the simulations. In the first configuration, the default WRF LU from a Moderate Resolution Imaging Spectroradiometer (MODIS) 30-s data product (modified 20-category IGBP-MODIS land use) is considered. This LU does not observe the existence of the mining facility. In the second configuration, WRF LU is overwritten with a MODIS 30-s data product (modified 21-category IGBP-MODIS land use) reflecting the existence of the mining facility. From Friedl et al. (2010), the following classifications have been used: pond (lake = 21), mine (barren = 16), plant and lodge (urban and built-up = 13), and other (grassland = 10). With this option, a one-dimensional mass and energy balance scheme for the pond is considered to model it as a lake (Subin et al. 2012; Gu et al. 2015). A depth of 50 m is considered, and the lake model accounts for 25 model layers including up to 5 layers for snow, 10 layers for water, and 10 layers for soil. LU changes in WRF are only considered for the topography case considering modifications with the SRTM-lidar topography datasets. The aerodynamic roughness length throughout the five domains is also appropriately forced by specifying the land-use classification. Figure 4 shows the two LU configurations for WRF simulations.

Fig. 4.
Fig. 4.

The two land-use configurations for WRF simulations; mine (barren = 16), pond (lake =21), plant and lodge (urban and built-up = 13), and other (grassland = 10); lakes are considered as class 21 during simulation but are not color coded.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

4) Planetary boundary layer parameterization

The PBL parameterization scheme implemented in the model plays a decisive role in its accuracy. Researchers have performed sensitivity tests of the WRF Model and found that the Yonsei University (YSU) scheme shows similar performance or improvement when compared with other PBL schemes of WRF, particularly in accounting for nonlocal and entrainment transport processes (Zhang and Zheng 2004; Hu et al. 2010; Hari Prasad et al. 2017; Fernández-González et al. 2018). Hence, the YSU scheme is applied to the domains with horizontal grid spacings of greater than 500 m. For domains with horizontal grid spacings of less than 500 m, the large-eddy simulation (LES) WRF configuration (WRF-LES) is used assuming that the resolved turbulent and terrain-driven eddies are mainly responsible for the bulk of the vertical and horizontal transfer of momentum, heat, water vapor, and atmospheric constituents within the PBL (Talbot et al. 2012; Chu et al. 2014).

5) Grid spacing

One of the main objectives of refining horizontal and vertical grid spacings is to reach a grid configuration that will be adequate to explicitly resolve atmospheric dynamics within the PBL (Aligo et al. 2009; Pontoppidan et al. 2017). Hence, the grid spacing is changed both in horizontal and vertical directions to find the optimum grid configuration in this study. The horizontal spacing for coarse, midscale, and fine grids in domain 5 are 0.18, 0.12, and 0.09 km, respectively, to effectively represent the resolution of the mine topography. The default 45 vertical levels are considered as coarse vertical grid spacing, whereas 90 and 120 vertical levels are implemented to obtain midscale and fine vertical grid spacings, respectively. Moreover, the coarse, midscale, and fine vertical spacings provide 13, 17, and 21 vertical levels, respectively, below 2700-m elevation (approximately the daytime PBL height). The grid is stretched in the vertical direction at midaltitudes and is refined near the bottom and top of the domain. The vertical grid spacing above the ground surface is 60 m.1 Grid-spacing details for all domains are presented in Table 1.

6) Evaluation

The experimental datasets from sodar are used to evaluate the WRF simulations at heights above 200 m above ground level (above the upper limit of the TANAB measurements), and the other instruments are used at heights at or below 200 m above ground level (i.e., near surface). A quantitative comparison between the WRF Model and experimental observations is performed by determining the bias and root-mean-square error (RMSE) defined by
bias=i=1n(MiOi)nand
RMSE=i=1n(MiOi)2n.
Here, Mi and Oi represent the WRF results and experimental observations, respectively, on an hourly basis for all three days of analysis and n = 72 represents the sample size. Because wind direction is a circular variable, differences of wind direction between WRF and the observations are reported as a positive number less than 180° by calculating the mean absolute error (MAE) (instead of bias) defined by Fernández-González et al. (2018):
MAE=i=0n|MiOi|n.
MAE for wind direction is calculated by excluding data when the measured wind speed is less than 2 m s−1 since under low wind conditions it is difficult to measure wind direction accurately.
To find a confidence interval for the range of the difference in bias (MAE for wind direction) values the t statistic is used. This is also known as a two-sample statistical estimation test that is used for a given atmospheric variable when running the model under different configurations. This inferential test quantifies the range for the change in bias (or MAE) and therefore quantifies any model improvement or deterioration as a result of changing model configurations. If x¯1 and s12, and x¯2 and s22 are the average (here bias or MAE) and standard deviation (here RMSE) of independent samples of size n1 and n2, respectively, an approximate (1 − α) × 100% confidence interval for μ1μ2, that is, the difference between the averages (here biases or MAEs), is
(x¯1x¯2)tα/2s12n1+s22n2<μ1μ2<(x¯1x¯2)+tα/2s12n1+s22n2,
where tα/2 is the t value with degrees of freedom
ν=(s12n1+s22n2)2(s12/n1)2n11+(s22/n2)2n21,
leaving an area α/2 to the right of a t probability distribution function (Walpole et al. 2002).
To study if the model predicts the temporal variation of different variables, the Pearson correlation coefficient (PCC) is also calculated as
PCC=i=1n(MiM¯)(OiO¯)i=1n(MiM¯)2i=1n(OiO¯)2,
where M¯ and O¯ represent the averages of WRF results and observations, respectively. A PCC close to 1 shows that the model can predict the temporal variation of different atmospheric variables well, while a value close to zero or negative indicate no correlation or anticorrelation, respectively (Walpole et al. 2002).

3. Results and discussion

The comparison between WRF simulations in domain 5 and the experimental observations on 18, 24, and 30 May 2018 is evaluated. Effects of topography and land-use modifications are considered first. Subsequently, the effects of horizontal/vertical grid spacings are discussed. For statistical analysis, both hourly statistics and overall statistics are calculated. For hourly statistics, the average value of a given variable for the three simulation dates and observations are considered for every hour, that is, the average of three numbers is calculated for every hour. For the overall statistics, all hourly values are considered for a given variable, that is, the bias (or MAE), RMSE, and PCC are calculated for 72 pairs of numbers. Also, if more than one instrument of a kind is present (e.g., three 2D and two 3D ultrasonic anemometers measured wind speed at 10 m above ground), the overall statistics are averaged among the number of instruments to report bias (or MAE), RMSE, and PCC. When the effects of topography and land use are discussed, hourly statistics and contour plots of near-surface-level variables for selected variables are shown. For brevity, hourly statistics and contour plots are not shown when the effects of grid spacing are studied. For these cases, only the overall statistics are reported.

Before detailed discussions of model configurations, selected profiles of wind speed and potential temperature as measured by TANAB can be investigated in Fig. 5. In the early morning from 0600 to 1000 mountain standard time (MST; =UTC − 7 h) 18 May 2018, wind speeds measured at all altitudes were less than 2 m s−1 indicative of calm conditions in the mine. In the early afternoon from 1200 to 1400 MST 24 May 2018 evidence of higher wind speed is found at altitudes 40–60 m above ground on the slopes to the east side of the tailings pond. These are upslope winds given that the measured local wind direction is from the south. In the late afternoon and early evening from 1400 to 2000 MST 30 May 2018 evidence of higher wind speed is found at altitudes 10–40 m above ground in the same location. These are downslope winds given that the measured local wind direction is from the north. The presence of this jet near the surface is in agreement with drainage flows upstream of the depression measured by Lehner et al. (2016) in the Arizona’s Meteor Crater experiment (their Fig. 7) and downslope flows measured by Clements et al. (2003) in Utah’s Peter Sinks (their Fig. 12). Controlled by topography, land cover, soil moisture, radiative exchange, local shading, and surface energy budget, such slope flows are hypothesized to form under fair weather conditions as a result of heating of atmospheric layers during daytime and cooling during nighttime (Hari Prasad et al. 2017). The potential temperature profile measured by TANAB in the early morning from 0400 to 0600 MST 18 May 2018 shows two thermally stable inversion layers, one forming just above the mine surface at the bottom and another above grade level (z ~ 100 m), with an almost isothermal layer in between. The isothermal layer is hypothesized to form due to air circulations and mixing inside the mine, radiative transfer, and air incursions across the depression edge (Whiteman et al. 2008). During the thermally stable conditions, similar temperature profiles were observed by Whiteman et al. (2008) in Arizona’s Meteor Crater experiment (their Fig. 10), Lehner et al. (2016) (their Fig. 8), and Clements et al. (2003) (their Figs. 5, 8, and 9). At later hours, TANAB observed development of a thermally unstable potential temperature profile both inside the mine and on the east side of the tailings pond, without a multilayer structure, in agreement with the aforementioned studies.

Fig. 5.
Fig. 5.

Selected vertical profiles of (a) mean wind speed and (b) mean potential temperature measured by the TANAB in the mine (on 18 and 24 May 2018) and on the east side of the pond (on 30 May 2018), grouped in 2-h time intervals; the vertical axis shows height above ground level; each mean was calculated over a 5-min sampling period.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

a. Effects of topography and land use

For this investigation, midhorizontal/vertical grid spacing WRF simulations are run to minimize the computational expense. The bias (or MAE), RMSE, and PCC of different variables are shown in Table 2. Moreover, the two-sample estimation tests among different cases are presented in Table 3.

Table 2.

Effects of topography and land-use changes on atmospheric dynamics; table shows the bias (MAE for wind direction), RMSE, and Pearson’s correlation coefficient (PCC) of different atmospheric variables for comparison between the WRF Model and field observations.

Table 2.
Table 3.

Effects of topography and land-use changes on atmospheric dynamics showing the two-sample estimation test (α = 0.1) among different cases: a confidence interval for difference in bias (MAE for wind direction) is reported for pairs of model configurations. Strictly positive confidence intervals are shown in boldface type.

Table 3.

1) Wind speed

Figure 6 shows the near-surface horizontal wind speed and direction in domain 5 for different types of topography and land-use options. It should be noted that the TANAB system is considered as a near-surface measurement because it only probes the atmospheric boundary layer up to low altitudes typically under 200 m.

Fig. 6.
Fig. 6.

Effects of topography and land-use changes on wind speed near the surface; the instruments were set up at 10 m except for the TANAB; the average value for the three dates at every hour is shown for both the model and the observations.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

It is observed that there is no major difference of wind speed predictions when comparing to 2D/3D ultrasonic anemometers at 10 m although at midnight and early morning the WRF case with SRTM-lidar topography, modified land use, and lake model marginally shows better agreement with the experimental observations (the PCC increases from 0.035 to 0.208). The reported biases and RMSEs here are in agreement with a study by Mughal et al. (2017) reporting a bias and RMSE of −0.29 and 1.2 m s−1, respectively, and another study by Fernández-González et al. (2018) reporting a bias and RMSE of 1.26 and 1.77 m s−1, respectively, using observations for other complex terrains. However, here the model shows the improvement of wind speed prediction with SRTM-lidar topography, modified land use, and lake model, when compared with TANAB and sodar observations with model biases reduced from 1.10 to 0.08 m s−1 and 1.04 to 0.50 m s−1, respectively. Corresponding confidence intervals in bias reduction are (0.46–1.58) and (0.30–1.60) m s−1 (α = 0.1), respectively. The PCC for the same cases also increases from 0.007 to 0.256 and from 0.262 to 0.430, respectively, suggesting more successful simulation of the temporal variation of wind speed when high-resolution topography, high-resolution land use, and lake modeling are considered. The reported PCCs in this study are lower than findings of Fernández-González et al. (2018) who demonstrate PCCs greater than 0.88 for wind-resource assessment over complex terrain of a mountain range. The difference in PCCs is likely due to the diversity of land-use classes and their abrupt changes in the present study. As the variations of wind speed at low altitudes depend on the surface roughness and atmospheric stability (Cerlini et al. 2005), the WRF case without the mine topography cannot perform accurate simulations. Moreover, in convective hours when the atmosphere is unstable, WRF shows lower bias than the thermally stable conditions at late night and early morning, in agreement with findings of Storm et al. (2009).

Figure 7 shows the contour plots of wind speed at 10 m for different types of topography and land-use options on 18 May 2018 during thermally stable (0200 MST) and unstable (1400 MST) conditions. There are some horizontal circulations (change in wind direction) over the mine cavity at 0200 MST in most of the WRF simulations accounting for the mine terrain. The winds are also slower in that area compared to the surroundings. Similar behavior was also reported by Clements et al. (2003) (their Fig. 11), that reported a cool and calm inversion layer at a bottom of an enclosed pit with wind direction changes compared to atmospheric layers above the inversion layer. During thermally stable (0200 MST) condition the temperature of the pond is higher than the surrounding land surfaces and the slightly warmer pond temperatures drive an airflow mechanism in cases accounting for land-use modification and lake model. However, during thermally unstable (1400 MST) condition, the temperature of the pond is lower than the surrounding land surfaces and the winds over the pond are slower in cases accounting for land-use modification and lake model.

Fig. 7.
Fig. 7.

Effects of topography and land-use changes on 10-m wind vector magnitude and direction at 0200 and 1400 MST 18 May 2018; black lines show mine and pond perimeters.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

2) Wind direction

Figure 8 shows the average values of absolute error for wind direction in domain 5 near-surface for different types of topography and land-use options. Overall, there are no major differences in wind direction among different WRF simulations. The MAE for wind direction does not reduce given the model configurations in topography and land use. It must be noted that the weather station inside the mine did not measure wind direction at 10 m, so the horizontal circulation of winds in the mine could not be confirmed experimentally, but the change in wind direction was expected given the observations of Clements et al. (2003) (their Fig. 11). There are higher average absolute errors for wind direction in thermally stable conditions, especially in early morning hours, compared to thermally unstable conditions. In addition, sometimes the average absolute error for wind direction is high during early morning and early evening when the atmospheric stability alternates from stable to unstable or unstable to stable conditions. Similar findings were reported by Talbot et al. (2012) who compared WRF-LES simulations with near-surface observations and noted higher errors for wind speed and direction at nighttime during the thermally stable conditions. In that study, the differences between the simulations and observations were attributed to the forcing datasets and a wide range of thermal stability conditions encountered in real cases, as opposed to turbulence parameterizations in the model, since lower errors were observed in idealized simulations.

Fig. 8.
Fig. 8.

Effects of topography and land-use changes on the absolute error of wind direction at 10 m; the average value of the absolute error of wind direction for the three dates at every hour is shown.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

3) Temperature and the turbulent sensible heat flux

Figure 9 shows the time series for the temperature at 2 m in domain 5 for different types of topography and land use. Measurements at 2 m show an overall low bias within ±0.5 K and RMSE under 2.5 K in predicting the temperature with PCC values generally greater than 0.9, regardless of topography and land-use configurations. WRF shows reasonable agreement with observations in both the thermally stable and unstable conditions. This is in agreement with findings of Cheng et al. (2013) who reported bias and RMSE of lower than −0.6 and 2 K, respectively, for different choices of land-use datasets, while higher resolution and more accurate land-use datasets reduced the error statistics. This is also in agreement with findings of Hari Prasad et al. (2017) who reported bias and RMSE of lower than −0.6 and 1.3 K, respectively. However, here the model shows the improvement of potential temperature prediction with SRTM-lidar topography, modified land use, and lake model when compared with TANAB observations with the model bias reduced from 0.98 to 0.32 K. The corresponding confidence interval in bias reduction is (0.01–1.31) K (α = 0.1). In addition, the PCC is greater than 0.9. He et al. (2017) also found that updated land surface information improved WRF’s performance in terms of daily average temperatures, and the RMSE was decreased by 7%.

Fig. 9.
Fig. 9.

Effects of topography and land-use changes on temperature at 2 m; the average value for the three dates at every hour is shown for both the model and the observations.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

The spatial distribution of potential temperature at 2 m both during thermally stable (0200 MST) and unstable (1400 MST) conditions on 18 May 2018 is shown in Fig. 10. Examining contour plots of temperature at 2 m provides a convenient way to understand the impact of altered topography and land use on meteorological conditions. Two mechanisms are considered to explain the observed effects: net radiation heat transfer between the Earth surface and the sky, and the heat capacity of the Earth surface. The land is always emitting longwave radiation upward to the atmosphere and space, but during the day this loss of longwave radiation from the surface is more than offset by incoming shortwave (solar) radiation. Given the low heat capacity of the land, this results in warm surface temperatures during the day and cool surface temperatures during the night. For water bodies, given the higher heat capacity, the temperature variation exhibits a lower amplitude diurnal cycle. This results in warmer water temperatures at night and cooler water temperatures during the day compared to the surrounding land surface temperatures (Dudhia 1989). Hence, in thermally stable hours (0200 MST), the pond surface exhibits a warmer temperature relative to the surrounding areas [and vice versa during the thermally unstable hours (1400 MST)]. However, the WRF case with GTOPO and SRTM-lidar topographies alone cannot capture this phenomenon. When mine topography and land use are accounted for by the model, examining the contour plot for thermally stable hours (0200 MST), a reduction in near-surface air temperature inside the mine can be noticed relative to the nearby grade-level areas. The reduction in temperature is about 5 K, in agreement with sinkholes with high sky view factors (~0.9 or alternatively width-to-depth aspect ratios greater than 5) reported by Whiteman et al. (2004), with sinkhole saddle–floor temperature differences in the range 5–10 K at the same hour (D0 and D1 sinkholes in their Fig. 5). Such differential temperatures can have profound effects on atmospheric dynamics over the mining facility.

Fig. 10.
Fig. 10.

Effects of topography and land-use changes on 2-m potential temperature at 0200 and 1400 MST 18 May 2018; black lines show mine and pond perimeters.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

The effect of topography and land use on turbulent sensible heat flux at 10 m is shown in Fig. 11. One of the major advantages of using land-use modifications and lake model is to predict the turbulent sensible heat flux more accurately with the model bias reduced from 45.7 to 17.3 W m−2. The corresponding confidence interval for bias reduction is (6.94–49.9) W m−2 (α = 0.1). WRF shows smaller variations of turbulent sensible heat flux during thermally stable conditions while the bias is higher during thermally unstable conditions. The PCC for all cases is greater than 0.8 showing the success of the model to predict the diurnal cycle of turbulent sensible heat flux. WRF-LES with fine grid spacing and high-resolution topography and land-use configurations has shown advantages over PBL schemes on coarse grid spacings when simulating turbulent sensible heat flux as well as the turbulent latent heat flux, particularly in predicting small-magnitude fluxes during thermally stable conditions of nighttime (Talbot et al. 2012).

Fig. 11.
Fig. 11.

Effects of topography and land-use changes on turbulent sensible heat flux at 10 m; the average value for the three dates at every hour is shown for both the model and the observations.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

4) Relative humidity

Relative humidity depends on temperature, pressure, and the water content of the atmosphere. Figure 12 shows the relative humidity at 2 m in domain 5 for different types of topography and land use. Overall, all WRF simulations exhibit the bias and RMSE around 1% and 10%, respectively. This is also in agreement with findings of Hari Prasad et al. (2017) who reported bias and RMSE of lower than 0.8% and 4.0%, respectively. All simulations exhibit a high PCC greater than about 0.9.

Fig. 12.
Fig. 12.

Effects of topography and land-use changes on relative humidity at 2 m; the average value for the three dates at every hour is shown for both the model and the observations.

Citation: Journal of Applied Meteorology and Climatology 59, 4; 10.1175/JAMC-D-19-0213.1

b. Effects of grid spacing

On the basis of the results in the previous section, only the case with SRTM-lidar topography, land-use change, and lake model is considered for this sensitivity study because overall the confidence intervals for bias reductions were maximized for these configurations. The bias (MAE), RMSE, and PCC of different variables for this investigation are shown in Table 4. The two-sample test was also employed in this case and is shown in Table 5. Note that the two-sample test was only employed for three pairs of grid configurations.

Table 4.

Effects of grid configuration showing the bias (MAE for wind direction), RMSE, and PCC of different atmospheric variables for comparison between WRF Model and field observations.

Table 4.
Table 5.

Effects of grid configuration changes on atmospheric dynamics showing the two-sample estimation test (α = 0.1) among different cases: a confidence interval for difference in bias (MAE for wind direction) is reported for pairs of model configurations. Strictly positive confidence intervals are shown in boldface type. CH-CV: coarse horizontal–coarse vertical, MH-MV: midscale horizontal–midscale vertical, and FH-FV: fine horizontal–fine vertical.

Table 5.

One of the main objectives of adjusting the horizontal and vertical grid spacings is to reach a configuration that will reduce the model biases the most. As far as wind speed at 10 m is concerned, there is a minor change in the confidence intervals for bias reductions, suggesting model improvements associated with refining horizontal or vertical grid spacings. For temperature, relative humidity, and turbulent sensible heat flux, there is no major change in the confidence intervals for bias reductions associated with refining or coarsening horizontal or vertical grid spacings. However, the optimum configuration to result in the highest confidence interval for bias reduction in wind speed in comparison to measurements of TANAB is the midscale horizontal and midscale vertical grid spacings. This will result in a confidence interval for bias reduction of (2.28–4.18) m s−1 (α = 0.1) in wind speed predictions near the surface up to 200-m altitude. This corresponds to a model bias reduction from 3.31 to 0.08 m s−1. Comparison with the sodar, however, results in the highest confidence interval for bias reduction in wind speed of (0.35–1.49) m s−1 associated with fine horizontal and fine vertical grid spacings although the PCC was degraded at these grid spacings. This corresponds to a model bias reduction from 0.80 to −0.11 m s−1. Carvalho et al. (2012), Talbot et al. (2012), Mirocha et al. (2016), and Fernández-González et al. (2018) have reported similar findings where refining grid spacings improved wind speed predictions up to a limit beyond which no further improvement was noticed. In addition, when using WRF-LES, Talbot et al. (2012) found that refining the grid spacing in the horizontal direction is more important than refining the grid in the vertical direction to reduce the model bias and RMSE. The effect of grid spacings on relative humidity and turbulent sensible heat flux is marginal although midhorizontal and midvertical or any grid spacings finer than this improve the turbulent sensible heat flux prediction. Except for the fine horizontal and fine vertical grid spacings, and for most variables, PCC values appear to be in the same range without being influenced by changing the grid spacing, suggesting that the model predicts the diurnal cycle of most variables with the same level of success given any grid spacing studied here. Overall, the grid configuration with a horizontal grid size of 0.12 or 0.09 km in the smallest domain and implementation of 90 vertical levels should be sufficient to provide accurate predictions in comparison to experimental observations at the mining facility.

4. Summary and conclusions

The sensitivity tests of the WRF Model are performed to investigate distinct meteorological conditions over complex topography of a remote open-pit mining facility in northern Canada in May 2018 under fair weather conditions without synoptic events. Three days in May 2018 were chosen to run the model. The sensitivity study varied the topography, land use, and horizontal and vertical grid spacings to arrive at the best configuration for running WRF while minimizing modeling biases in comparison to field observations. Comparisons of model results to observations are performed in both near-surface and higher elevations above ground using various meteorological instruments and weather stations including 2D/3D ultrasonic anemometers, TANAB, sodar, cup anemometers, temperature, and relative humidity sensors. The default surface topography in the model is modified using the SRTM 1-s dataset in combination with local lidar observations, while land-use change is implemented from a MODIS 30-s data product. The YSU PBL scheme is used where horizontal grid spacing is greater than 500 m, and WRF-LES is used where horizontal grid spacing is less than 500 m.

The model exhibits a better performance when topography, land-use change, and lake modeling are accounted for relative to the alternative of default topography and land use with no lake modeling. As a result, the model biases reduce from 1.10 to 0.08 m s−1, from 1.04 to 0.50 m s−1, from 0.98 to 0.32 K, and from 45.7 to 17.3 W m−2, for near-surface wind speed, boundary layer wind speed, near-surface potential temperature, and turbulent sensible heat flux, respectively. The corresponding confidence intervals for bias reduction are (0.46–1.58) m s−1, (0.30–1.60) m s−1, (0.01–1.31) K, and (6.94–49.9) W m−2 (α = 0.1), for near-surface wind speed, boundary layer wind speed, near-surface potential temperature, and turbulent sensible heat flux, respectively. Hence, incorporating high-resolution topography, modifying land use, and implementing a lake model should be considered for more accurate WRF simulations. As far as grid configuration is concerned, refining model horizontal and vertical grid spacings results in bias reductions from 3.31 to 0.08 and from 0.80 to −0.11 m s−1, for near-surface and boundary layer wind speeds, respectively. The corresponding confidence intervals for bias reduction are (2.28–4.18) and (0.35–1.49) m s−1 (α = 0.1), associated with near-surface and boundary layer wind speeds, respectively.

The simulations also depict meteorological phenomena in agreement with previous studies of enclosed depressions of Earth. There are some horizontal circulations over the mine cavity under thermally stable conditions in most of the WRF simulations accounting for the mine terrain. The winds are also slower in that area relative to the surroundings, in agreement with reported cool and calm inversion layers at the bottom of enclosed pits with wind direction changes relative to atmospheric layers above the inversion layer. During thermally stable conditions, the temperature of the pond is higher than the surrounding land surfaces and the slightly warmer pond temperatures drive an airflow mechanism in cases accounting for land-use modification and lake model. However, during thermally unstable conditions, the temperature of the pond is lower than the surrounding land surfaces and the winds over the pond are slower in cases accounting for land-use modification and lake model.

From the findings in this study, it is suggested to run WRF simulations with modified topography, modified land use, and lake modeling for meteorological predictions over complex open-pit mining facilities. The grid configuration with a horizontal grid size of 0.12 or 0.09 km in the smallest domain and implementation of 90 vertical levels for all domains should be sufficient to provide accurate predictions in comparison to experimental measurements. These configurations result in statistically notable reductions in model biases in comparison with observations. This improvement cannot be overlooked and can serve more accurate meteorological modeling toward operational plume dispersion models attempting to quantify area-fugitive emission fluxes from open-pit mining facilities. This is a first step toward this goal.

The limitations of this study include lack of more rigorous observations and area-distributed measurements with larger spatial coverage. There is also a limitation of measuring temperature profiles to evaluate the performance of WRF at higher elevations within PBL. In addition, ideally more than three days of simulations should be considered if more computational power is available to provide enough statistical sampling. Future research should also investigate seasonal variations of different atmospheric variables to understand the meteorological conditions over the complex open-pit mining facility more comprehensively under fair weather and synoptic events. Future research should employ tracer dispersion modeling of a plume of greenhouse gas at the mining facility to understand meteorological implications over a complex topography with land-use alterations in area-fugitive emissions. The current WRF results can also be potentially used as boundary conditions for fine grid spacing computational fluid dynamics simulations.

Acknowledgments

The Tethered and Navigated Air Blimp was partially developed by the assistance of Denis Clement, Jason Dorssers, Katharine McNair, James Stock, Darian Vyriotes, Amanda Pinto, and Phillip Labarge. The authors thank Andrew F. Byerlay for designing and constructing the tether reel system for TANAB. The authors are indebted to Steve Nyman, Chris Duiker, Peter Purvis, Manuela Racki, Jeffrey Defoe, Joanne Ryks, Ryan Smith, James Bracken, and Samantha French at the University of Guelph, who helped with the campaign logistics. Special credit is directed toward Amanda Sawlor, Datev Dodkelian, Esra Mohamed, Di Cheng, Randy Regan, Margaret Love, and Angela Vuk at the University of Guelph for administrative support. The computational platforms were set up with the assistance of IT staff Jeff Madge, Joel Best, and Matthew Kent at the University of Guelph. Technical discussions with John D. Wilson and Thomas Flesch at the University of Alberta are highly appreciated. In-kind technical support for this work was provided by Rowan Williams Davies and Irwin, Inc. (RWDI). This work was supported by the University of Guelph, Ed McBean philanthropic fund; the Discovery Grant program (401231) from the Natural Sciences and Engineering Research Council (NSERC) of Canada; Government of Ontario through the Ontario Centres of Excellence (OCE) under the Alberta-Ontario Innovation Program (AOIP) (053450); and Emission Reduction Alberta (ERA) (053498). OCE is a member of the Ontario Network of Entrepreneurs (ONE).

Data availability statement: The Atmospheric Innovations Research (AIR) Laboratory at the University of Guelph may provide the confidential supporting field data via the authorization of the data owners. For access, contact Principal Investigator Amir A. Aliabadi (aliabadi@uoguelph.ca).

REFERENCES

  • Aliabadi, A. A., M. Moradi, D. Clement, W. D. Lubitz, and B. Gharabaghi, 2019: Flow and temperature dynamics in an urban canyon under a comprehensive set of wind directions, wind speeds, and thermal stability conditions. Environ. Fluid Mech., 19, 81109, https://doi.org/10.1007/s10652-018-9606-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Aligo, E. A., W. A. Gallus Jr., and M. Segal, 2009: On the impact of WRF Model vertical grid resolution on Midwest summer rainfall forecasts. Wea. Forecasting, 24, 575594, https://doi.org/10.1175/2008WAF2007101.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Awan, N. K., H. Truhetz, and A. Gobiet, 2011: Parameterization-induced error characteristics of MM5 and WRF operated in climate mode over the Alpine region: An ensemble-based analysis. J. Climate, 24, 31073123, https://doi.org/10.1175/2011JCLI3674.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Byerlay, R. A. E., M. K. Nambiar, A. Nazem, M. R. Nahian, M. Biglarbegian, and A. A. Aliabadi, 2020: Measurement of land surface temperature from oblique angle airborne thermal camera observations. Int. J. Remote Sens., 41, 31193146, https://doi.org/10.1080/01431161.2019.1699672.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Carvalho, D., A. Rocha, M. Gómez-Gesteira, and C. Santos, 2012: A sensitivity study of the WRF model in wind simulation for an area of high wind energy. Environ. Modell. Software, 33, 2334, https://doi.org/10.1016/j.envsoft.2012.01.019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cerlini, P. B., K. A. Emanuel, and E. Todini, 2005: Orographic effects on convective precipitation and space-time rainfall variability: Preliminary results. Hydrol. Earth Syst. Sci., 9, 285299, https://doi.org/10.5194/hess-9-285-2005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cheng, F.-Y., Y.-C. Hsu, P.-L. Lin, and T.-H. Lin, 2013: Investigation of the effects of different land use and land cover patterns on mesoscale meteorological simulations in the Taiwan area. J. Appl. Meteor. Climatol., 52, 570587, https://doi.org/10.1175/JAMC-D-12-0109.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cheng, W. Y., and W. J. Steenburgh, 2005: Evaluation of surface sensible weather forecasts by the WRF and the eta models over the western United States. Wea. Forecasting, 20, 812821, https://doi.org/10.1175/WAF885.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chu, X., L. Xue, B. Geerts, R. Rasmussen, and D. Breed, 2014: A case study of radar observations and WRF LES simulations of the impact of ground-based glaciogenic seeding on orographic clouds and precipitation. Part I: Observations and model validations. J. Appl. Meteor. Climatol., 53, 22642286, https://doi.org/10.1175/JAMC-D-14-0017.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Clements, C. B., C. D. Whiteman, and J. D. Horel, 2003: Cold-air-pool structure and evolution in a mountain basin: Peter Sinks, Utah. J. Appl. Meteor., 42, 752768, https://doi.org/10.1175/1520-0450(2003)042<0752:CSAEIA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • De Meij, A., and J. F. Vinuesa, 2014: Impact of SRTM and Corine land cover data on meteorological parameters using WRF. Atmos. Res., 143, 351370, https://doi.org/10.1016/j.atmosres.2014.03.004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Diagne, M., M. David, J. Boland, N. Schmutz, and P. Lauret, 2014: Post-processing of solar irradiance forecasts from WRF model at Reunion Island. Sol. Energy, 105, 99108, https://doi.org/10.1016/j.solener.2014.03.016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dudhia, J., 1989: Numerical study of convection observed during the Winter Monsoon Experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46, 30773107, https://doi.org/10.1175/1520-0469(1989)046<3077:NSOCOD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fernández-González, S., M. L. Martín, E. García-Ortega, A. Merino, J. Lorenzana, J. L. Sánchez, F. Valero, and J. S. Rodrigo, 2018: Sensitivity analysis of the WRF Model: Wind-resource assessment for complex terrain. J. Appl. Meteor. Climatol., 57, 733753, https://doi.org/10.1175/JAMC-D-17-0121.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flesch, T. K., J. D. Wilson, and E. Yee, 1995: Backward-time Lagrangian stochastic dispersion models and their application to estimate gaseous emissions. J. Appl. Meteor., 34, 13201332, https://doi.org/10.1175/1520-0450(1995)034<1320:BTLSDM>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flesch, T. K., J. D. Wilson, L. A. Harper, and B. P. Crenna, 2005: Estimating gas emissions from a farm with an inverse-dispersion technique. Atmos. Environ., 39, 48634874, https://doi.org/10.1016/j.atmosenv.2005.04.032.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flesch, T. K., J. D. Wilson, L. A. Harper, R. W. Todd, and N. A. Cole, 2007: Determining ammonia emissions from a cattle feedlot with an inverse dispersion technique. Agric. For. Meteor., 144, 139155, https://doi.org/10.1016/j.agrformet.2007.02.006.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Friedl, M. A., D. Sulla-Menashe, B. Tan, A. Schneider, N. Ramankutty, A. Sibley, and X. Huang, 2010: MODIS Collection 5 global land cover: Algorithm refinements and characterization of new datasets. Remote Sens. Environ., 114, 168182, https://doi.org/10.1016/j.rse.2009.08.016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gilliam, R. C., and J. E. Pleim, 2010: Performance assessment of new land surface and planetary boundary layer physics in the WRF-ARW. J. Appl. Meteor. Climatol., 49, 760774, https://doi.org/10.1175/2009JAMC2126.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gu, H., J. Jin, Y. Wu, M. B. Ek, and Z. M. Subin, 2015: Calibration and validation of lake surface temperature simulations with the coupled WRF-lake model. Climatic Change, 129, 471483, https://doi.org/10.1007/s10584-013-0978-y.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hari Prasad, K. B. R. R., C. V. Srinivas, T. N. Rao, C. V. Naidu, and R. Baskaran, 2017: Performance of WRF in simulating terrain induced flows and atmospheric boundary layer characteristics over the tropical station Gadanki. Atmos. Res., 185, 101117, https://doi.org/10.1016/j.atmosres.2016.10.020.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • He, J. J., Y. Yu, L. J. Yu, N. Liu, and S. P. Zhao, 2017: Impacts of uncertainty in land surface information on simulated surface temperature and precipitation over China. Int. J. Climatol., 37, 829847, https://doi.org/10.1002/joc.5041.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., 2012: Orographic effects on precipitating clouds. Rev. Geophys., 50, RG1001, https://doi.org/10.1029/2011RG000365.

  • Hu, N., T. K. Flesch, J. D. Wilson, V. S. Baron, and J. A. Basarab, 2016: Refining an inverse dispersion method to quantify gas sources on rolling terrain. Agric. For. Meteor., 225, 17, https://doi.org/10.1016/j.agrformet.2016.05.007.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hu, X.-M., J. W. Nielsen-Gammon, and F. Zhang, 2010: Evaluation of three planetary boundary layer schemes in the WRF Model. J. Appl. Meteor. Climatol., 49, 18311844, https://doi.org/10.1175/2010JAMC2432.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jiménez, P. A., and J. Dudhia, 2012: Improving the representation of resolved and unresolved topographic effects on surface wind in the WRF Model. J. Appl. Meteor. Climatol., 51, 300316, https://doi.org/10.1175/JAMC-D-11-084.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lehner, M., and Coauthors, 2016: The METCRAX II field experiment: A study of downslope windstorm-type flows in Arizona’s Meteor Crater. Bull. Amer. Meteor. Soc., 97, 217235, https://doi.org/10.1175/BAMS-D-14-00238.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liggio, J., and Coauthors, 2019: Measured Canadian oil sands CO2 emissions are higher than estimates made using internationally recommended methods. Nat. Commun., 10, 1863, https://doi.org/10.1038/s41467-019-09714-9.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lin, Y.-L., R. D. Farley, and H. D. Orville, 1983: Bulk parameterization of the snow field in a cloud model. J. Climate Appl. Meteor., 22, 10651092, https://doi.org/10.1175/1520-0450(1983)022<1065:BPOTSF>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lorente-Plazas, R., P. A. Jiménez, J. Dudhia, and J. P. Montávez, 2016: Evaluating and improving the impact of the atmospheric stability and orography on surface winds in the WRF Model. Mon. Wea. Rev., 144, 26852693, https://doi.org/10.1175/MWR-D-15-0449.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mass, C. F., D. Ovens, K. Westrick, and B. A. Colle, 2002: Does increasing horizontal resolution produce more skillful forecasts? The results of two years of real-time numerical weather prediction over the Pacific Northwest. Bull. Amer. Meteor. Soc., 83, 407430, https://doi.org/10.1175/1520-0477(2002)083<0407:DIHRPM>2.3.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mirocha, J. D., M. D. Simpson, J. D. Fast, L. K. Berg, and R. Baskett, 2016: Investigation of boundary-layer wind predictions during nocturnal low-level jet events using the Weather Research and Forecasting model. Wind Energy, 19, 739762, https://doi.org/10.1002/we.1862.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mughal, M. O., M. Lynch, F. Yu, B. McGann, F. Jeanneret, and J. Sutton, 2017: Wind modelling, validation and sensitivity study using Weather Research and Forecasting model in complex terrain. Environ. Modell. Software, 90, 107125, https://doi.org/10.1016/j.envsoft.2017.01.009.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nossent, J., P. Elsen, and W. Bauwens, 2011: Sobol’ sensitivity analysis of a complex environmental model. Environ. Modell. Software, 26, 15151525, https://doi.org/10.1016/j.envsoft.2011.08.010.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pontoppidan, M., J. Reuder, S. Mayer, and E. W. Kolstad, 2017: Downscaling an intense precipitation event in complex terrain: The importance of high grid resolution. Tellus, 69A, 1271561, https://doi.org/10.1080/16000870.2016.1271561.

    • Search Google Scholar
    • Export Citation
  • Powers, J. G., and Coauthors, 2017: The Weather Research and Forecasting Model: Overview, system efforts, and future directions. Bull. Amer. Meteor. Soc., 98, 17171737, https://doi.org/10.1175/BAMS-D-15-00308.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Qiu, X., I. Cheng, F. Yang, E. Horb, L. Zhang, and T. Harner, 2018: Emissions databases for polycyclic aromatic compounds in the Canadian Athabasca oil sands region—Development using current knowledge and evaluation with passive sampling and air dispersion modelling data. Atmos. Chem. Phys., 18, 34573467, https://doi.org/10.5194/acp-18-3457-2018.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Saide, P. E., G. R. Carmichael, S. N. Spak, L. Gallardo, A. E. Osses, M. A. Mena-Carrasco, and M. Pagowski, 2011: Forecasting urban PM10 and PM2.5 pollution episodes in very stable nocturnal conditions and complex terrain using WRF-Chem CO tracer model. Atmos. Environ., 45, 27692780, https://doi.org/10.1016/j.atmosenv.2011.02.001.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Storm, B., J. Dudhia, S. Basu, A. Swift, and I. Giammanco, 2009: Evaluation of the Weather Research and Forecasting Model on forecasting low-level jets: Implications for wind energy. Wind Energy, 12, 8190, https://doi.org/10.1002/we.288.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Subin, Z. M., W. J. Riley, and D. Mironov, 2012: An improved lake model for climate simulations: Model structure, evaluation, and sensitivity analyses in CESM1. J. Adv. Model. Earth Syst., 4, M02001, https://doi.org/10.1029/2011MS000072.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Talbot, C., E. Bou-Zeid, and J. Smith, 2012: Nested mesoscale large-eddy simulations with WRF: Performance in real test cases. J. Hydrometeor., 13, 14211441, https://doi.org/10.1175/JHM-D-11-048.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Taylor, C. M., E. F. Lambin, N. Stephenne, R. J. Harding, and R. L. Essery, 2002: The influence of land use change on climate in the Sahel. J. Climate, 15, 36153629, https://doi.org/10.1175/1520-0442(2002)015<3615:TIOLUC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Walpole, R. E., R. E. Myers, S. L. Myers, and K. Ye, 2002: Probability & Statistics for Engineers & Scientists. 7th ed. Prentice Hall, 730 pp.

    • Search Google Scholar
    • Export Citation
  • Whiteman, C. D., T. Haiden, B. Pospichal, S. Eisenbach, and R. Steinacker, 2004: Minimum temperatures, diurnal temperature ranges, and temperature inversions in limestone sinkholes of different sizes and shapes. J. Appl. Meteor., 43, 12241236, https://doi.org/10.1175/1520-0450(2004)043<1224:MTDTRA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Whiteman, C. D., and Coauthors, 2008: METCRAX 2006: Meteorological experiments in Arizona’s Meteor Crater. Bull. Amer. Meteor. Soc., 89, 16651680, https://doi.org/10.1175/2008BAMS2574.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Xu, L., H. Liu, Q. Du, and L. Wang, 2016: Evaluation of the WRF-lake model over a highland freshwater lake in southwest China. J. Geophys. Res. Atmos., 121, 13 98914 005, https://doi.org/10.1002/2016JD025396.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yan, H., and W. A. Gallus Jr., 2016: An evaluation of QPF from the WRF, NAM, and GFS models using multiple verification methods over a small domain. Wea. Forecasting, 31, 13631379, https://doi.org/10.1175/WAF-D-16-0020.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Yáñez-Morroni, G., J. Gironás, M. Caneo, R. Delgado, and R. Garreaud, 2018: Using the Weather Research and Forecasting (WRF) model for precipitation forecasting in an Andean region with complex topography. Atmosphere, 9, 304, https://doi.org/10.3390/atmos9080304.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, C., H. Lin, M. Chen, and L. Yang, 2014: Scale matching of multiscale digital elevation model (DEM) data and the Weather Research and Forecasting (WRF) model: A case study of meteorological simulation in Hong Kong. Arab. J. Geosci., 7, 22152223, https://doi.org/10.1007/s12517-014-1273-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, D.-L., and W.-Z. Zheng, 2004: Diurnal cycles of surface winds and temperatures as simulated by five boundary layer parameterizations. J. Appl. Meteor., 43, 157169, https://doi.org/10.1175/1520-0450(2004)043<0157:DCOSWA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zheng, Y., K. Alapaty, J. A. Herwehe, A. D. Del Genio, and D. Niyogi, 2016: Improving high-resolution weather forecasts using the Weather Research and Forecasting (WRF) Model with an updated Kain–Fritsch scheme. Mon. Wea. Rev., 144, 833860, https://doi.org/10.1175/MWR-D-15-0005.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhu, R., and D. Xu, 2004: Multi-scale turbulent planetary boundary layer parameterization in mesoscale numerical simulation. Quart. J. Appl. Meteor., 15, 543555.

    • Search Google Scholar
    • Export Citation
1

Eta levels η(z) = [P(z) − PTop]/(PSurfacePTop) for coarse vertical grid spacing are 1.000, 0.991, 0.980, 0.965, 0.947, 0.923, 0.897, 0.868, 0.840, 0.811, 0.782, 0.741, 0.689, 0.640, 0.594, 0.550, 0.509, 0.470, 0.433, 0.399, 0.366, 0.335, 0.306, 0.279, 0.254, 0.230, 0.207, 0.186, 0.166, 0.148, 0.131, 0.115, 0.100, 0.086, 0.0741, 0.063, 0.053, 0.043, 0.035, 0.027, 0.020, 0.014, 0.008, 0.003, and 0.000. Eta levels for midscale vertical grid spacing are 1.000, 0.991, 0.980, 0.965, 0.947, 0.925, 0.897, 0.876, 0.863, 0.850, 0.837, 0.818, 0.793, 0.769, 0.746, 0.722, 0.700, 0.678, 0.656, 0.635, 0.615, 0.595, 0.576, 0.557, 0.538, 0.520, 0.503, 0.485, 0.469, 0.452, 0.436, 0.421, 0.406, 0.391, 0.377, 0.363, 0.349, 0.336, 0.323, 0.311, 0.298, 0.287, 0.275, 0.264, 0.253, 0.242, 0.232, 0.222, 0.212, 0.202, 0.193, 0.184, 0.175, 0.167, 0.159, 0.150, 0.143, 0.135, 0.128, 0.121, 0.114, 0.107, 0.101, 0.095, 0.089, 0.083, 0.078, 0.073, 0.068, 0.063, 0.059, 0.054, 0.050, 0.046, 0.042, 0.038, 0.035, 0.031, 0.028, 0.025, 0.022, 0.019, 0.016, 0.013, 0.011, 0.008, 0.006, 0.003, 0.001, and 0.000. Eta levels for fine vertical grid spacing are 1.000, 0.991, 0.980, 0.965, 0.947, 0.925, 0.897, 0.878, 0.869, 0.859, 0.849, 0.835, 0.817, 0.799, 0.781, 0.764, 0.747, 0.730, 0.713, 0.697, 0.681, 0.665, 0.650, 0.635, 0.620, 0.605, 0.591, 0.577, 0.563, 0.549, 0.536, 0.523, 0.510, 0.497, 0.485, 0.472, 0.460, 0.449, 0.437, 0.426, 0.415, 0.404, 0.393, 0.383, 0.372, 0.362, 0.352, 0.343, 0.333, 0.324, 0.315, 0.306, 0.297, 0.288, 0.280, 0.271, 0.263, 0.255, 0.247, 0.240, 0.232, 0.225, 0.218, 0.210, 0.204, 0.197, 0.190, 0.184, 0.177, 0.171, 0.165, 0.159, 0.153, 0.147, 0.142, 0.136, 0.131, 0.126, 0.120, 0.115, 0.111, 0.106, 0.101, 0.097, 0.092, 0.088, 0.084, 0.080, 0.076, 0.073, 0.069, 0.065, 0.062, 0.059, 0.056, 0.052, 0.049, 0.046, 0.044, 0.041, 0.038, 0.035, 0.033, 0.030, 0.028, 0.026, 0.023, 0.021, 0.019, 0.017, 0.015, 0.013, 0.011, 0.009, 0.007, 0.006, 0.004, 0.002, and 0.001.

Save