a geophysical grid, constructed using gridded terrain and land cover data (obtained from GeoGratis Government of Canada); and

Transcription

1 1.0 INTRODUCTION The following is a summary of model inputs and odour modelling results conducted for the purpose of assessing potential odour impacts from a private organics management facility located at 2760 Marron Valley Road, Kaleden, British Columbia (hereafter referred to as the Site ). Odour modelling was conducted using CALPUFF, an advanced air modelling software system recommended by the British Columbia Ministry of Environment (BC MOE). 2.0 MODEL INPUTS AND ASSUMPTIONS 2.1 Meteorology The air dispersion model CALPUFF contains a diagnostic meteorological processor, CALMET, which creates a three-dimensional meteorological field over the spatial extent of the model. The data produced by CALMET is used by CALPUFF in its dispersion and plume transport calculations. Inputs to CALMET include the following: a geophysical grid, constructed using gridded terrain and land cover data (obtained from GeoGratis Government of Canada); and a combination of prognostic (three-dimensional meso-scale model called MM5) meteorological data and hourly surface observations obtained from Environment Canada and BC MOE meteorological stations. When CALMET is run in no-observations mode (using only MM5), the surface station observations provide a validation of the CALMET meteorology, in particular winds, to ensure representativeness. As MM5 is a meso-scale regional model, the grid used as input to CALMET is downscaled in three steps from a 32 km resolution grid to a 4 km grid and downscaled again within CALMET to the CALPUFF grid size (250 m). It is not expected that the meteorological time series in CALMET will exactly reproduce observed conditions on an hour by hour basis at any particular grid point, however it is expected to be representative of the general conditions over a given year. Table 2.1 summarizes the meteorological inputs to CALMET used in the odour modelling and mapping exercise for the Site. Table 2.1: CALMET Inputs and Metadata Parameter Surface Stations Upper Air Soundings Usage None None Prognostic Data 4 km resolution MM5 (2012 & 2013) Meteorological Grid 40 km (east-west) x 30 km (north-south) at 250 m 2 Grid Centre 312,000 m, 5,473,000 m, UTM Zone 11 Vertical Cells (Cell Face Heights) Terrain Data Land Use Data 10 (0 m, 20 m, 40 m, 80 m, 160 m, 320 m, 640 m, 1,200 m, 2,000 m, 3,000 m, 4,000 m) CDN DEM 15 min (082e03-e06, e11,e12) GeoBase Land Cover circa 2000-Vector (082e) 1

2 CALMET uses energy balance algorithms to compute hourly gridded fields of various micrometeorological parameters which, combined with the gridded vertical temperature profiles contained in MM5, produces the three-dimensional atmospheric stability field that drives the dispersion model. The inputs to the energy balance are derived from the geophysical (terrain and land cover) grid. As land cover characteristics in the region vary with season (e.g., snow-covered cropland has a much higher albedo than during the summer), in accordance with British Columbia Air Quality Dispersion Modelling Guideline (BCMOE 2015), five seasonal CALMET files were created using the recommended geophysical parameters for each land cover category for each seasonal period. The date ranges used to define each season are listed in Table 2.2. Year-to-year variability will undoubtedly occur, however, this temporal approximation was used to simplify modelling based on Environment Canada Climate Normals for the Okanagan-Similkameen region. The modelled year was Table 2.2: Geophysical Property Seasonality Season Date Range Winter 1 October 16 December 10 Winter 2 December 11 February 29 Transitional Spring March 1 May 31 Summer June 1 September 15 Fall September 15 October Meteorological Validations To assess the representativeness of the CALMET meteorological grid, validation assessments were conducted on the CALMET wind field (are the expected terrain effects and temporal patterns present?) and the vertical atmospheric stability (does CALMET capture winter inversions and capture the expected seasonal/diurnal and spatial mixing height patterns?) Winds Figure 2.1 shows annual wind roses extracted from the CALMET data at five locations in the Kaleden model grid. The wind roses illustrates predominant wind directions as the length of a particular directional plot is representative of the frequency of occurrence of winds from that direction. Each directional plot is further broken up into the relative frequency of occurrence of wind speeds from that direction navy blue representing wind speeds between 0.3 and 1 m/s for instance. Figure 2.1 shows good agreement with the expected flow pattern: aligned north-south within the Okanagan Valley with drainage winds aligned in the orientation of the local valley terrain. At the site, the predominant winds are from the south-southwest, due to the orientation of the mountain pass southwest of the Site. Northwesterly/northerly winds are of secondary predominance, as a result of flows originating from the plateau to the west and northerly winds which funnel through the Marron River valley from the north. The temporal nature of the predominant wind patterns are illustrated in Figure 2.2 as seasonal wind roses of the modelled CALMET winds for winter (DJF) and summer (JJA). Correlating with the seasonal regional flow patterns, southwesterly winds are the common winter pattern while summer winds are heavily influenced by the diurnal localized valley/mountain breeze pattern. During the daytime, heating of the air within the Okanagan Valley causing it to rise, flowing up-valley, resulting in easterly winds in the vicinity of the Site. At night, cooler, denser air over the plateau flows down-valley through the various mountain passes, resulting in northwesterly winds in the vicinity of the Site. 2

3 Figure 2.1: CALMET-Modelled Winds through Kaleden Domain Annual (Jan 2012 Mar 2013) Wind Roses Figures 2.3 through 2.5 show snapshots of the typical winter and summer wind patterns respectively. Figure 2.3 shows the predominant southerly flow through the region during winter with valley-steered winds in the vicinity of the site. Figure 2.4 illustrates the typical summer nighttime pattern with drainage winds from the west flowing through the Marron River valley from the north and over mountain slopes from the northwest. Figure 2.5 illustrates the typical summer daytime valley breeze condition occurring under clear skies, as ground heating within the Okanagan Valley enhances convection (the vertical movement of air), resulting in calmer winds at the valley floor and up-valley flows onto the mountain ridges and the plateau to the west. 3

4 Figure 2.2: CALMET-Modelled Winds at 2760 Marron Valley Road Winter (DJF, left) and Summer (JJA, right) Figure 2.3 Snapshot of CALMET-Modelled Winter Winds (Jan. 1, 19:00) 4

5 Figure 2.4 Snapshot of CALMET-Modelled Summer Nighttime Winds (Jul. 4, 23:00) Figure 2.5 Snapshot of CALMET-Modelled Summer Daytime Winds (Jul. 5, 13:00) 5

6 Vertical Temperature Profiles Vertical temperature profiles illustrate atmospheric stability in the CALMET data and indicate the presence of inversions. The normal condition in the lower atmosphere is air temperature decreasing with altitude. Simplified, under this condition, the lower atmosphere can generally be considered as neutral or unstable, meaning air that is uplifted, due to heating from the ground for instance, has a tendency to rise and odours released at the surface more readily disperse. The rate of the decrease with height is indicative of the overall degree of atmospheric stability. When temperatures increase with height in the lower atmosphere, an inversion is present. The height of the inversion is indicated by the inflection point of the profile that is the point where temperature begins to decrease with altitude. During inversion conditions, vertical movement is supressed and the dispersion of odours is diminished, usually leading to higher concentrations at ground level. Inversions are typical during winter, particularly in valley scenarios where colder denser air sinks to the valley floor and is contained by valley walls. Inversions also occur overnight in the absence of ground heating. Figure 2.6 shows diurnal vertical temperature profiles extracted from the CALMET data, in six hour intervals for both January 1 (left) and July 1 (right) in the modelled year 2012 at the location of the Site. The January plots show the presence of weak inversions in the modelled data. Overnight (0:00, red and 6:00, orange), the inversion height is near the ground (the model default is 50 m). After sunrise, as the sun heats the ground, which in turn warms the air near the ground, the inversion breaks up (called fumigation which can lead to some of the highest odour concentrations at the surface) and the atmosphere becomes more unstable. The plot for 12:00 (green), shows the normal condition of decreasing temperature with altitude. At 18:00 (blue), ground heating subsides after sunset (~17:00) and the inversion condition begins to build. In contrast, the diurnal profiles for July 1 indicate a normal, unstable atmosphere with the exception of 0:00 where stable conditions are present near the surface due to the absence of heating Mixing Height The atmospheric mixing height can be defined as the top of the layer in the lower atmosphere, within which an emitted species, in this case odour, is readily mixed through turbulence and convective processes. When the mixing height is low, higher ground-level concentrations will generally be predicted. Mixing height is greatly influenced by wind speed (higher wind speeds induce greater turbulence), incident solar radiation (ground heating of the air near the surface induces convection and vertical movement of air), terrain (uneven terrain is more conducive to turbulence) and geophysical characteristics (e.g. surface roughness) of the surface. Figure 2.7 shows the spatial pattern to mixing heights as grid cell values for a night and day scenario in January. Figure 2.8 shows the same for July. In Figure 2.7, stable (inversion) conditions (very low mixing heights) are present in the valley bottoms, indicated by the purple shaded grid cells. Along the valley ridges, winds blowing over uneven terrain and forested areas (high surface roughness) induces turbulence, resulting in locally elevated mixing heights, indicated by the green and yellow shaded cells. During the day, incident solar radiation is insufficient to warm the cooler dense air trapped at the valley bottoms, and mixing heights remain quite low, in the range of 100 m to 200 m, resulting in stable, stagnant conditions which persist over long durations. 6

7 Figure 2.6: CALMET-Modelled Diurnal Vertical Temperature Profiles at 2760 Marron Valley Road (January 1, left, July 1, right) 7

8 Figure 2. 7: CALMET-Modelled Mixing Height January 2, 2012 (00:00, left & 16:00, right) Figure 2.8: CALMET-Modelled Mixing Height July 5, 2012 (00:00, left & 13:00, right) In Figure 2.8, the contrast in mixing heights between day and night is quite evident. Overnight, cool, denser air fills the valley resulting in stable conditions. Terrain-induced turbulence is observed along the valley walls as cooler air from the plateau flows into the valley. In the daytime, strong solar heating induces convection throughout the area, resulting in mixing heights exceeding 2,000 m, with the exception being over Okanagan and Vaseux Lake where the heat flux interactions are much lower than over land due to the properties of water. 8

9 Figure 2.9: CALMET-Modelled Mixing Heights at 2760 Marron Valley Road Winter 2012 (JF, blue) and Summer 2012 (JJA, red) Figure 2.9 plots hourly mixing heights extracted from the CALMET data at the location of the Site. The upper plot (blue) shows mixing heights for January and February while the lower plot (red) shows mixing heights for summer (June through August). As expected, with lower sun angles, less incident solar radiation and frequent inversions, winter mixing heights are generally much lower than in summer. The diurnal pattern (lower mixing heights overnight) is quite evident in both plots but is much more evident in the summer due to the strong heating differential in the region. 2.2 Emission Factors The site layout was based on the membrane cover layout for the Summerland Regional Facility Feasibility Assessment. This layout was used to define the boundaries of the odour sources for this modelling analysis and is included in Appendix A as Figure 1. Odourous air from composting is managed through a GORE cover, which is a breathable membrane. Odourous air from receiving area (which is inside a covered fabric building) is managed through a biofilter. The largest odour source is the receiving area biofilter. Other odour sources include the composting piles, storage area, curing piles, and receiving area. All odour sources were considered area sources, and assumed to occur homogeneously over the entirety of the area source. Another source of odour emissions occurs from movement of materials, i.e. pile building, turning, and moving. This type of activity occurs intermittently and were assigned a diurnal variation based on the expected times of day the activity is to be performed. Such activities are expected to occur daily at the Site over a one- to two-hour period, however since the activity may occur at any time during the operational hours of the facility in the morning or in the afternoon, odour emissions were assumed in the model to occur between to representing a time of day 9

10 when vertical mixing is generally highest and between to when, during the winter, the mixing height is approaching its night time minimum, thus resulting in higher concentrations closer to the ground. This is a somewhat conservative approach since the activity may only be occurring over a portion of a single hour rather than four, may not take place every day, and peak odour emission would only occur during and immediately following the activity and decay in the hour following. It should be noted that odour emissions produced from pile building and moving are inconsequential compared to that produced from the biofilters which emit odour continuously. Emission factors used for this model are presented in Table 2.3. Table 2.3: Emission Factors Emission Source Emission Factor 1 Time Total Area Type Release Height Emission Factor Source Storage OU/m 2 s Continuous (24 h) 12,000 m 2 Area 3 m Phase 1 Odour Modelling Report Screening OU/m 2 s Continuous (24 h) 900 m 2 Area 1 m Phase 1 Odour Modelling Report Receiving Areas OU/m 2 s Continuous (24 h) 2,112 m 2 Area 1 m Phase 1 Odour Modelling Report Biofilters (by Receiving Areas) OU/m 2 s Continuous (24 h) 352 m 2 Area 1 m Phase 1 Odour Modelling Report Composting 0.2 OU/m 2 s Continuous (24 h) 11,550 m 2 Area 3 m Phase 1 Odour Modelling Report Compost Pile Building 0.44 OU/m 2 s 10:00-12:00, 14:00-16: m 2 Area 3 m Phase 1 Odour Modelling Report Compost Pile Moving 0.47 OU/m 2 s 10:00-12:00, 14:00-16: m 2 Area 3 m Phase 1 Odour Modelling Report 2.3 CALPUFF Settings and Assumptions The CALPUFF model input settings were assigned with consideration to the recommendations in Table 7.1 Recommendations for Key CALPUFF (Version: 6.42, Level: ) Model Options in Input Group 2 and 12 in Guidelines for Air Quality Dispersion Modelling in British Columbia (BCMOE 2015). Generally, default model settings were used. Since the area of interest is in the near-field (within km of the source), dispersion coefficients were internally calculated using micrometeorological variables (MDISP = 2) based on estimates of the crosswind and vertical components of turbulence based on similarity theory and the land cover type. The probability distribution function (PDF) was used for dispersion under convective conditions (MPDF = 1) which explicitly accounts for the differences in the distribution and strengths of up and down drafts within the convective boundary layer, reporting the average between the two. By using these two settings, AERMOD-type dispersion is simulated (generally accepted as better-predicting in the near-field than CALPUFF), while also providing the benefit of a puff model and allowing for the effects of complex terrain. The ground level receptor grid spacing was 250 m over the entire grid. Additional 15 km (N-S) x 4 km (W-E) receptor grids were added at 10 m, 20 m, 40 m, 60 m and 80 m above ground, centered over the Golden Mile facility to observe the pattern of concentrations with height. The modelled stack is in close proximity to the dome-shaped main facility building which can produce wake-effect turbulence that can draw the odour plume down to the surface. 10

11 The main building dimensions were approximated from engineering drawings and downwash was included as a model option. 3.0 RESULTS Since the time step of the meteorological data is one-hour, CALPUFF can only output one-hour averaged predictions of odour concentration. However, since odour perception is on a much shorter scale, an averaging timescalar must be applied to assess shorter-term peak concentrations due to plume meandering within the hourly period. Hourly odour concentrations are scaled to a 10-minute averaging period using Equation 1. =. (1) Pursuant to Equation 1, to is the 60 minute averaging time, tp is the short-term averaging time (10 minutes) and Co and Cp are the respective peak concentrations (BC MOE). The scalar when converting from hourly to 10-minute average concentrations equates to Odour Units An Odour Unit (OU) is a way of quantifying odours through the use of an odour panel that consists of a group of people with calibrated noses. The definition of an OU is based on the proportion of odour panel members that can detect the smell of a substance. One OU represents the concentration of a particular substance when 50% of the odour panel can detect the odour. This is called the perception threshold 1. At this point, although an odour may be detected, it is not distinct enough to be able to identify the type of odour. The OU scale is based on dilutions, as shown in the following figure. As the number of odour units increase, more people can detect the odour, and the intensity of the odour increases. Five OU is considered a faint odour and 10 OU is considered a distinct odour (the point when some people can identify the type of odour, or its potential source) Odours and VOCs: Measurement, Regulation and Control Techniques (2009). Kassel University Press. 11

12 Figure 3.1: Odour Unit Scale There are currently no guidelines for odour limits for composting facilities in British Columbia, however, some wastewater treatment facilities have imposed odour limits. For example, the standard in Metro Vancouver is no more than 5 OU at the property line. In other jurisdictions, the guideline is to have no detectable odour at the property line. At the Ogogrow facility in Vernon, BC, the limit is 50 OU at the property line. 3.2 Odour Maps Odour maps are included as part of Appendix A. Odour modelling results are presented as three different plots: Maximum Odour Concentrations The maximum predicted 10-minute odour concentration at each receptor point over the course of the modelled year. This is displayed as a contour plot showing the maximum predicted 10-minute averaged odour concentration at every ground level receptor point over the entire one-year simulation (8,784 hours) as a blue gradient (light to dark). The 1 OU contour is white. The highest levels >10 OU are dark blue. The facility boundary is shown as a green outline. Hourly Exceedances >1 OU The number of hours over the course of the modelled year where an odour threshold of 1 OU was exceeded in a 10-minute averaged concentration. This is displayed as a contour plot showing the number of times the predicted 10-minute odour concentration exceeded 1 OU over the modelled year (2012) as an orange gradient (light to dark). The white contour line represents <20 exceedances per year. This would theoretically equate to 50% of the population being able to detect odour produced by the facility less than 0.2% of the time. The dark orange contour line represents >100 exceedances per year. Hourly Exceedances >5 OU The number of hours over the course of the modelled year where an odour threshold of 5 OU was exceeded in a 10-minute averaged concentration. This is displayed as a contour plot showing the number of times the predicted 10-minute odour concentration exceeded 5 OU over the modelled year (2012) as an orange gradient (light to dark). The white contour line represents <20 exceedances per year. 12

13 This would theoretically equate to when a faint odour is produced by the facility less than 0.2% of the time. The dark orange contour line represents >100 exceedances per year. 3.3 Results Summary The odour maps presented in Appendix A show: 1) the magnitude and spatial extent of maximum ground level odour, and 2) the number of exceedances of odour detection thresholds. Figure 2 shows the maximum predicted 10-minute averaged ground level concentration at all grid points over the entire simulation. It is a conglomerate of all modelled hours (8,784) and is not a snapshot at a given time. The figure shows the spatial extent of areas affected by transport of odour away from the proposed facility and the potential magnitude of odour impact. The highest levels (>10 OU) are predicted in the vicinity of the facility, confined within in the valley and mountain passes due in part to terrain confinement and in part to prevalent stagnant, stable conditions which lead to elevated odour concentrations at ground level. Downwind of the facility, ground level concentrations slightly exceeding 1 OU were also predicted through connecting valleys and passes as well as through the Okanagan Valley as downslope winds transport odour down the Marron Valley and into the main Okanagan Valley flow. The highest levels in the Okanagan Valley (4 6 OU) are predicted over Skaha Lake and Vaseux Lake, the result of generally lower mixing heights over water. Figures 3 and 4 show the number of hours with predicted exceedances of 1 OU and 5 OU, respectively. As expected, the majority of exceedances occur in the vicinity of the Site. Away from the Site, the number of hours with exceedances of 1 OU generally are below 20 per year, or 0.2% of the time. There are zero exceedances of 5 OU away from the Site. The following table summarizes the results of the odour mapping exercise based on the predicted maximum odour and number of hours of odour exceedances at a location approximately 400 m south of the property boundary (800 m south of the composting area) representing the resident that is closest in proximity to the Site ( , ), as shown in Figure 3.2. Table 3.1: Results Summary based on Closest Receptor Point Location Maximum Predicted 10-min Odour Odour Exceedance >1 OU (hours per year) Odour Exceedance >5 OU (hours per year) , OU

14 Figure 3.2: Location of Nearest Discrete Receptor ( , ) Attachments: Appendix A (4 figures) 14

15 APPENDIX A Figure 1 Figure 2 Figure 3 Figure 4 Site Layout Maximum Predicted Ground Level Odour Concentration (Over a Sustained 10-Minute Period) within the Course of 1 Year (Current Composting Operations) Number of Hours with Exceedances of 1 OU (Detectable Odour by 50% of the Population) within the Course of 1 Year Number of Hours with Exceedances of 5 OU (Detectable Odour by 50% of the Population) within the Course of 1 Year 15

16 Marron Valley Rd Bulking Agent Storage Biofilter Storage Receiving Area Biofilter Screening Receiving Area Q:\Vancouver\GIS\SOLID_WASTE\SWOP\SWOP _RDOS\Maps\SWOP _Fig01_Kaleden_Membrane.mxd modified 14/06/2016 by stephanie.leusink LEGEND Loader Movement Truck Movement Composting Screening and Storage Receiving Area Mech; Biofilter Parcel Boundary Mech Composting NOTES Base data source: Parcel boudaries for Penticton 1 provided by Canada Lands Digital Cadastral Data (downloaded June 14, 2016). Imagery provided by ESRI; DigitalGlobe (2010). Composting PROJECTION Mech FILE NO. SWOP _Fig01_Kaleden_Membrane.mxd STATUS OFFICE DATE ISSUED FOR REVIEW Tt EBA-VANC June 14, 2016 DATUM PROJECT NO. DWN CKD APVD REV SWM.SWOP SL MEZ BL 0 CLIENT PHASE 2 ODOUR MODELLING 2760 Marron Valley Road, Kaleden, BC Membrane Covered Aerated Static Pile Site Layout UTM Zone 11 NAD83 Scale: 1:3, Metres Figure Regional District of Okanagan-Similkameen

17 APPENDIX A FILE: SWM.SWOP JUNE 2016 ISSUED FOR REVIEW Facility Boundary Figure 2: Maximum Predicted Ground Level Odour Concentration (Over a Sustained 10-Minute Period) within the Course of 1 Year (Current Composting Operations) 1 App A Fig 2 -

18 APPENDIX A FILE: SWM.SWOP JUNE 2016 ISSUED FOR REVIEW Facility Boundary Figure 3: Number of Hours with Exceedances of 1 OU (Detectable Odour by 50% of the Population) within the Course of 1 Year 2 App A Fig 2 -

19 APPENDIX A FILE: SWM.SWOP JUNE 2016 ISSUED FOR REVIEW Facility Boundary Figure 4: Number of Hours with Exceedances of 5 OU (Detectable Odour by 50% of the Population) within the Course of 1 Year 3 App A Fig 2 -