Supplements

Supplementary figures, tables and descriptions to article “Validation of the PALM model system 6.0 in a real urban environment; case study on Prague-Dejvice, Czech Republic”

S1. Common figures

Figure S1. Location of the Prague-Dejvice quarter with delineation of outer (red line) and inner (blue line) domain. Vector geodata for the top left image was provided by ESRI (Europe NUTS 0 Boundaries). Administrative borders on top right images (The Data200 database), same as the base map of the Czech Republic at 1:10 000 for the city of Prague on bottom image, were provided by Czech Office for Surveying, Mapping and Cadastre (ČÚZK, 2020).

Figure S2. Digital elevation model of Prague-Dejvice with delineation of parent (red line) and child (blue line) domain. Digital elevation model was provided by Prague OpenData portal (Prague Geoportal, 2020), the base map of the Czech Republic at 1:10 000 for the city of Prague was provided by Czech Office for Surveying, Mapping and Cadastre (ČÚZK, 2020).

Figure S3. Urban Atlas 2012 (Copernicus Land Monitoring Service) land cover of Prague-Dejvice with delineation of parent (red line) and child (blue line) domain. The base map of the Czech Republic at 1:10 000 for the city of Prague was provided by Czech Office for Surveying, Mapping and Cadastre (ČÚZK, 2020).

Figure S4. Semi-automatically processed facade image (left), circular temperature measurement areas and transformation points are shown. Ground image processed by FLIR Tools (right), rectangular temperature measurement areas were selected manually.

Figure S5. Comparison of facade surface temperatures measured by the FLIR SC660 IR camera (squares and triangles) and TRSYS01 thermal sensor (solid and dashed line) at Sinkule house during the summer (A) and winter (B) campaign. GF (blue colour) and 1F (black colour) stands for ground floor and 1st floor measurements respectively.

Figure S7. Locations of measuring vehicles (MV) in Prague-Dejvice. MV positions in the study area are marked on Fig. 1.

Figure S8. Relative humidity (green) observed using drone measurements of vertical profiles on 28 November 2018. Drone position is marked on Fig. 1. Purple line shows drone height in time.

Figure S9. Air temperature (green) observed using drone measurements of vertical profiles on 28 November 2018. Drone position is marked on Fig. 1. Purple line shows drone height in time.

Figure S10. Locations of air quality background stations and selected meteorological stations in Prague. The ortophoto for the city of Prague was provided by Czech Office for Surveying, Mapping and Cadastre as WMS service (ČÚZK, 2020).

Figure S11. Example 3D view of the parent modelling domain with 10m resolution from south-west direction on 20 July 2018 at 13:00 UTC (14:00 CET). The colour scale represents the modelled surface temperature.

Figure S12. The top view to the parent modelling domain with 10m resolution. The borders of the grid boxes are shown with grey lines, the footprint of the discretized buldings is denoted by the red colour.

Figure S13. Geopotential height at 500 hPa in 6-hour timestep for the summer e2 and winter e3 episodes. Blue line is ERA-Interim reanalysis, red line is WRF simulation of 9km domain. Background layer uses the public domain shaded relief map from Shaded Relief (2020).

Figure S14. Geopotential height at 850 hPa in 6-hour timestep for the summer e2 and winter e3 episodes. Blue line is ERA-Interim reanalysis, red line is WRF simulation of 9km domain. Background layer uses the public domain shaded relief map from Shaded Relief (2020).

Figure S15. Vertical profile of potential temperature (solid line) and wind speed (dashed line) for the summer e1 episode from the weather balloon soundings at Praha-Libuš station (purple and red) and parent domain average from the PALM model simulation (yellow and blue).

Figure S16. Vertical profile of potential temperature (solid line) and wind speed (dashed line) for the winter e1 episode from the weather balloon soundings at Praha-Libuš station (purple and red) and parent domain average from the PALM model simulation (yellow and blue).

Figure S17. Vertical profile of potential temperature (solid line) and wind speed (dashed line) for the winter e2 episode from the weather balloon soundings at Praha-Libuš station (purple and red) and parent domain average from the PALM model simulation (yellow and blue).

Figure S18. Observations of NO_X (top) and PM₁₀ (bottom) from urban background stations closest to PALM domains (see Fig. S10) during summer e1 and e2 episodes (left) and winter e1, e2, and e3 episodes (right) are compared with the results of CAMx model. Mean and its 95% confidence interval in mean are shown.

Figure S19. Overview of the situation in the location 11_1_V; the aerial view of the modelled child domain from south-east. The color scale represents the modelled surface temperature on 5 December 2018 at given times (UTC), the red arrow shows area of the location 11_1_V. For technical reasons, the step times for the model views express minutes as decimal fractions of the hours.

Figure S21. Time series of wind speed and wind direction at the roof of the tallest building of the Faculty of Civil Engineering of the Czech Technical University for summer episode e1 (top), winter episode e1 (middle), and winter episode e2 (bottom). The graphs show wind speed and the boxes of arrows wind direction. The red colour represents the observations, the blue colour the PALM modelled values, green colour values from the WRF model, and the black line the values from the nearest synoptic station at Praha-Ruzyne. Thin dotted lines represent 10-minute averages and the thick solid lines 1-hour moving averages of wind speed. The arrows represent 2-hour averages of wind direction. PALM model results are taken from the child domain with 2 m horizontal resolution.

Figure S22. Results of the grid discretization experiment from Section 5.1.7. Time series of the surface temperature, shortwave irradiance, and net radiation on the south oriented wall in three heights above the pavement, averaged horizontally over the central third of the canyon. These values come from the two idealized PALM simulations of the west-east oriented street canyon. The simulation marked “parallel” has the grid aligned with the canyon, while the simulation marked “diagonal” has the grid rotated 45° with respect to the canyon, creating a step-wise discretization pattern.

Figure S23. Results of the grid discretization experiment from Section 5.1.7. Vertical profiles of the surface temperature, shortwave irradiance, and net radiation on the south oriented wall, averaged horizontally over the central third of the canyon, in selected three hours of the day. These values come from the two idealized PALM simulations of the west-east oriented street canyon. The simulation marked “parallel” has the grid aligned with the canyon, while the simulation marked “diagonal” has the grid rotated 45° with respect to the canyon, creating a step-wise discretization pattern.

S2. Common tables

Table S1. Measurement point locations with their coordinates and local climate zone (Stewart and Oke, 2012).
Name	Type	Longitude	Latitude	LCZ
FSv rooftop wind	anemometer	14.38900	50.10461	4
Jug. p. St. west	MV	14.39481	50.10359	2/5
Jug. p. St. east	MV	14.39526	50.10361	2/5
Orlík h. east	MV	14.39934	50.10198	2/5
Orlík h. west	MV	14.39914	50.10194	2/5
Bubeneč h. east	MV	14.39750	50.10505	2/5
Bubeneč. h. west	MV	14.39734	50.10497	2/5
Sinkule h.	MV	14.39015	50.10125	2/B
HF Sinkule h.	HF	14.39014	50.10151	2/5
HF Zelená St.	HF	14.39575	50.10692	2/5
IR 01	IR	14.39042	50.10160	2/5
IR 02	IR	14.39263	50.10239	4
IR 03	IR	14.39009	50.10348	2/E
IR 04	IR	14.38973	50.10470	4
IR 05	IR	14.39156	50.10446	B
IR 06-1, IR 06-2	IR	14.39498	50.10392	2
IR 06-3	IR	14.39506	50.10354	2
IR 07	IR	14.39617	50.10498	2/5
IR 08	IR	14.39733	50.10519	2/5
IR 09	IR	14.39775	50.10213	2/5
IR 10	IR	14.39505	50.10134	D
IR 11	IR	14.39071	50.10034	2/E
IR 12	IR	14.39014	50.10127	2/B

Abbreviations: LCZ = local climate zone; MV = monitoring vehicle; HF = heat flux observation; IR = observation with infrared camera.

Table S2. Schedule of the summer observation campaign.
Date	PALM validation episode	MV positions¹	MV at Sinkule house²	Mobile AQ³	IR camera	Heatflux⁴	FSv rooftop wind⁵	Traffic intensities (camera)
10.–11.07.2018		Bubeneč E + W	✓				✓
12.07.2018		Bubeneč E + W	✓	✓			✓
13.07.2018		Bubeneč E + W to Orlík E + W	✓				✓
14.–15.07.2018	e1	Orlík E + W	✓				✓	✓
16.07.2018	e1	Orlík E + W to Bubeneč E + Orlík E	✓				✓	Orlík W and Bubeneč E only
17.07.2018		Bubeneč E + Orlík E	✓				✓	Bubeneč E only
18.07.2018		Bubeneč E + Orlík E	✓	✓			✓	Bubeneč E only
19.07.2018	e2	Bubeneč E + Orlík E to Jug. p. E + W	✓	✓	✓	Sinkule 1F	✓	Bubeneč E and Jug. p. E only
20.–21.07.2018	e2	Jug. p. E + W	✓		✓	Sinkule 1F + G	✓	✓
22.–23.07.2018	e2	Jug. p. E + W	✓			Sinkule 1F + G	✓	✓
24.07.2018						Sinkule 1F + G	✓
25.07.–02.08.2018						Sinkule 1F + G
03.08.2018						Sinkule 1F + G to Zelená
04.–07.08.2018						Zelená

Abbreviations: MV = monitoring vehicle; IR = observation with infrared camera; AQ = air quality observation.

¹ Air quality (NO_X, NO₂, NO, O₃, SO₂, CO, PM₁₀, PM_2.5, and PM₁) and meteorology (wind speed and direction and turbulent flow characteristics measured by the METEK 3D ultrasonic anemometer, air temperature, relative humidity, global radiation, and atmospheric pressure. Data from the METEK anemometer had a 10-min resolution, other variables were available at 1-min resolution. E/W marks the position on the east/west side of street.
² As in ¹ but PM₁, PM_2.5 and turbulence characteristics were not measured; wind speed and direction were measured by the GILL 2D WindSonic anemometer.
³ Air quality (NO_X, NO₂, NO, O₃, PM₁₀, PM_2.5, and PM₁), meteorology (wind speed and direction, air temperature), and information on traffic. PM and meteorology measurements not available on 12 Jul.
⁴ 1F is location in height of first floor (±5 m), G is in ground floor (±1 m).
⁵ Wind speed and direction in 1-sec resolution.

Table S3. Schedule of winter observation campaign.
Date	PALM validation episode	MV positions¹	MV at Sinkule house²	Mobile AQ³	IR camera	Heatflux⁴	FSv rooftop wind⁵	Traffic intensities (camera)
23.11.2018		Bubeneč E + W	✓				✓	✓
24.–25.11.2018	e1	Bubeneč E + W	✓				✓	✓
26.11.2018	e1	Bubeneč E + W	✓	✓		Zelená	✓	✓
27.11.2018	e2	Bubeneč E + W to Orlík E + W	✓			Zelená	✓	✓
28.11.2018	e2	Orlík E + W	✓			Zelená	✓ +drone	✓
29.11.2018	e2	Orlík E + W	✓			Zelená	✓	✓
30.11.2018		Orlík E + Orlík W to Bubeneč E	✓			Zelená	✓	✓
01.–02.12.2018		Orlík E + Bubeneč E	✓				✓	✓
03.12.2018		Bubeneč E + Orlík E to Jug. p. E	✓			Sinkule 1F + G	✓	✓
04.12.2018	e3	Bubeneč E + Jug. p. E	✓	✓	✓	Sinkule 1F + G	✓	✓
05.12.2018	e3	Bubeneč E + Jug. p. E	✓		✓	Sinkule 1F + G	✓ +drone	✓
06.12.2018	e3	Bubeneč E + Jug. p. E	✓		✓	Sinkule 1F + G	✓	✓
07.12.2018		Jug. p. E + Bubeneč E to Jug. p. W	✓			Sinkule 1F + G	✓	✓
08.12.2018		Jug. p. E + W	✓			Sinkule 1F + G	✓	✓
09.–10.12.2018		Jug. p. E + W	✓			Sinkule 1F + G	✓	✓

Abbreviations: MV = monitoring vehicle; IR = observation with infrared camera; AQ = air quality observation.

¹ Air quality (NO_X, NO₂, NO, O₃, SO₂, CO, PM₁₀, PM_2.5, and PM₁) and meteorology (wind speed and direction and turbulent flow characteristics measured by the METEK 3D ultrasonic anemometer, air temperature, relative humidity, global radiation, and atmospheric pressure. Data from the METEK anemometer had a 10-min resolution, other variables were available at 1-min resolution. E/W marks the position on the east/west side of street.
² As in ¹ but PM₁, PM_2.5 and turbulence characteristics were not measured; wind speed and direction were measured by the GILL 2D WindSonic anemometer.
³ Air quality (NO_X, NO₂, NO, O₃, PM₁₀, PM_2.5, PM₁, and SO₂ - 26 Nov only), meteorology (wind speed and direction, air temperature, relative humidity - 4 Dec only), and information on traffic.
⁴ 1F is location in height of first floor (±5 m), G is in ground floor (±1 m).
⁵ Wind speed and direction in 1-sec resolution. Drone measured air temperature, relative humidity and number of particles.

Table S4. Meteorological stations used for WRF evaluation.
Code	WMO code	Name	Longitude	Latitude	Altitude (m ASL)
P1PKAR01	11519	Praha, Karlov	14.427800	50.069200	261
P1PLIB01	11520	Praha, Libuš	14.446900	50.007800	302
P1PRUZ01	11518	Praha, Ruzyně	14.255600	50.100300	364

Table S5. Air quality stations used for evaluation of the CAMx model.
Code	Name	Classification¹	Longitude	Latitude	Elevation (m)	Method² PM₁₀	Method² NO_X
AKOBA	Praha 8-Kobylisy	B/S/R	14.467578	50.122189	269	RADIO	CHLM
ASUCA	Praha 6-Suchdol	B/S/R	14.384639	50.126528	277	RADIO	not measured
AREPA	Praha 1-n. Republiky	B/U/C	14.429220	50.088065	190	RADIO	CHLM
ARIEA	Praha 2-Riegrovy sady	B/U/NR	14.442692	50.081483	256	RADIO	CHLM
ABREA	Praha 6-Brevnov	B/U/RN	14.380116	50.084385	300	RADIO	CHLM

¹ Station EoI (exchange of information) classification (97/101/EC): B - Background, S - Suburban, U - Urban, R - Regional, C - Comercial, NR - natural, residential, RN - residential, natural.
² Measurement methods: RADIO - radiometry; beta ray absorption; CHLM - chemiluminescence.

Table S6. Surface categories and parametres for the BSM model.
Category code	Surface						Volume			Description
Category code	Albedo	Emissivity	Roughness length (m)	Roughness length heat (m)	Volumetric thermal capacity (J K^-1 m^-3)	Thermal conduc-tivity (W m^-1 K^-1)	Thickness (m)	Volumetric thermal capacity (J K^-1 m^-3)	Thermal conduc-tivity (W m^-1 K^-1)	Surface	Storage
101	0.30	0.93	0.0002	0.0002	1,000,000	2.0	0.40	1,000,000	0.2	sheet metal	insulated bricks
102	0.40	0.93	0.0012	0.0012	1,360,000	0.5	0.40	1,000,000	0.2	bricks	insulated bricks
103	0.35	0.93	0.0010	0.0010	1,360,000	0.5	0.40	1,000,000	0.2	plain bricks	insulated bricks
104	0.40	0.93	0.0003	0.0003	1,200,000	0.3	0.40	1,000,000	0.2	plain paving	insulated bricks
105	0.35	0.93	0.0007	0.0007	1,200,000	0.3	0.40	1,000,000	0.2	plain tiling	insulated bricks
106	0.40	0.93	0.0009	0.0009	1,300,000	0.5	0.40	1,000,000	0.2	plaster	insulated bricks
107	0.40	0.93	0.0009	0.0009	1,000,000	0.2	0.40	600,000	0.1	insulating plaster	insulated blocks
108	0.40	0.93	0.0010	0.0010	1,080,000	0.5	0.40	1,000,000	0.2	plaster + bricks	insulated bricks
109	0.30	0.95	0.0050	0.0050	1,080,000	0.5	0.40	1,000,000	0.2	plaster + green wall	insulated bricks
110	0.40	0.93	0.0010	0.0010	1,190,000	0.4	0.40	1,000,000	0.2	plaster + tiling	insulated bricks
111	0.40	0.93	0.0007	0.0007	2,100,000	0.6	0.40	1,000,000	0.2	plaster + glass	insulated bricks
112	0.40	0.93	0.0007	0.0007	1,500,000	0.2	0.40	1,000,000	0.1	plastics	insulated structure
113	0.50	0.92	0.0007	0.0007	2,100,000	0.9	0.40	1,000,000	0.2	glass	bricks
114	0.50	0.92	0.0007	0.0007	2,100,000	0.9	0.35	1,008,000	0.5	glass	insulation
115	0.50	0.92	0.0007	0.0007	2,100,000	0.6	0.40	600,000	0.15	glass + plaster	fired blocks
116	0.40	0.93	0.0008	0.0008	2,450,000	0.6	0.35	25,000	0.035	plaster + glass	insulation
117	0.40	0.93	0.0011	0.0011	1,000,000	0.3	0.40	1,000,000	0.1	stucco	insulated bricks
118	0.50	0.92	0.0007	0.0007	2,100,000	0.5	0.35	100,000	0.1	glass	steel-concrete structure
119	0.50	0.92	0.0007	0.0007	1,500,000	0.2	0.35	25,000	0.035	plastics layer	air
120	0.50	0.92	0.0010	0.0010	1,600,000	1.0	0.35	960,000	0.2	concrete	concrete+insulation
121	0.50	0.92	0.0100	0.0100	2,100,000	0.9	0.35	100,000	0.2	metro-bus-tram stops	air
122	0.50	0.92	0.0100	0.0100	2,000,000	0.2	0.35	100,000	0.2	shelters	air
123	0.50	0.92	0.0050	0.0050	2,000,000	0.2	0.35	2,000,000	0.2	wood	wood
201	0.50	0.93	0.0010	0.0010	1,144,000	0.5	0.30	1,144,000	0.3	sheet metal	softwood
202	0.60	0.70	0.0100	0.0100	1,500,000	0.7	0.35	1,144,000	0.5	concrete/fired tile	concrete/brick/wood
203	0.20	0.93	0.0010	0.0010	956,800	0.5	0.30	1,144,000	0.3	asphalt belt	softwood
204	0.30	0.96	0.0100	0.0100	1,500,000	1.0	0.45	1,500,000	1.0	grass	soil
205	0.30	0.96	0.0500	0.0500	1,500,000	1.0	0.45	1,500,000	1.0	grass+shrubs	soil
206	0.50	0.92	0.0007	0.0007	1,300,000	1.0	0.35	1,512,000	1.0	glass	air
207	0.66	0.59	0.0100	0.0100	1,125,000	1.6	1.00	1,328,000	1.5	pebble gravel	concrete
208	0.55	0.91	0.1000	0.1000	4,187,000	0.6	1.00	6,699,200	0.6	equipments	concrete
209	0.64	0.85	0.0010	0.0010	1,840,000	1.6	1.00	1,360,000	1.6	concrete	concrete
210	0.40	0.93	0.0007	0.0007	2,000,000	0.7	0.35	1,512,000	1.0	plastics	bricks

Table S7. Assignment of evaluation surface points to surface categories.
Category	Evaluation points	Section
pavements	01-2_H1, 02-1_H1, 02-1_H2, 02-2_H1, 02-3_H2, 02-4_H2, 03-1_H1, 03-2_H1, 03-2_H2, 03-2_H3, 04-1_H1, 05-1_H2, 06-1_H1, 06-1_H2, 06-2_H1, 06-2_H2, 07-1_H1, 07-1_H2, 07-2_H1, 07-2_H2, 10-1_H1, 10-1_H2, 10-1_H3, 10-2_H2, 10-2_H5, 11-1_H1, 11-2_H1, 12-1_H1, 12-1_H2	4.3.2
streets	01-1_H1, 01-1_H2, 02-2_H2, 06-1_H3, 06-2_H3, 06-3_H1, 07-1_H3, 07-1_H4, 07-2_H3, 08-1_H1, 08-1_H2, 08-2_H1, 08-2_H2, 09-1_H1, 09-1_H2, 09-1_H3, 09-2_H1, 09-2_H2, 09-2_H3, 09-2_H4, 10-2_H3, 10-2_H4, 11-1_H2, 11-2_H2	4.3.2
grass	02-3_H1, 02-4_H1, 04-1_H1, 05-1_H1, 05-1_H3, 05-1_H4, 05-2_H1, 06-3_H2, 10-3_H1	4.3.2
wall (traditional building)	01-1_V1, 01-1_V2, 01-2-B-1..3, 02-2_V2, 02-2_V3, 03-2_V1, 03-2_V2, 06-2_V1..5, 06-4_V1..4, 07-1_V1..4, 07-2_V1..5, 08-1_V2, 08-2_V2..4, 09-1_V1..5, 09-2_V1..7, 11-1_V1..5, 12-1_V1..3	4.3.3
wall (contemporary office building)	02-1_V1, 02-3_V1, 02-3_V2, 02-4_V1, 02-4_V2, 04-2_V1, 04-2_V2, 04-4_V1, 05-2_V1	4.3.3
wall (glass like surface building)	03-1_V1, 03-1_V2, 04-1_V2, 04-3_V1, 04-3_V2, 05-1_V1, 06-1_V1, 06-1_V2, 11-2_V1..3	4.3.3
plant canopy affected surface	02-1_H2, 02-2_H1, 02-2_H2, 04-1_H4, 05-1_H1, 05-1_H4, 05-2_H2, 06-2_H2, 08-1_V2, 08-1_H1, 08-1_H2, 08-2_V2..4, 08-2_V1, 08-2_V2, 09-1_V1..5, 09-1_V2, 09-2_V7, 09-2_H2, 09-2_H4, 12-1_H2	4.3.4

Table S8. Comparison of 1-h average PM₁₀ concentrations measured in the street canyons with CAMx and PALM results for the same location.
	Summer episodes		Winter episodes		All episodes
	PALM	CAMx	PALM	CAMx	PALM	CAMx
N	225	225	380	376	605	601
mean obs (μg m^-3)	23.8	23.8	30.7	30.8	28.1	28.2
mean mod (μg m^-3)	21.6	8.5	41.3	25.2	34.0	18.9
standard deviation obs (μg m^-3)	18.0	18.0	16.1	16.1	17.1	17.2
standard deviation mod (μg m^-3)	13.8	5.0	24.1	13.0	22.9	13.4
FB	-0.1	-0.9	0.3	-0.2	0.2	-0.4
NMSE	0.5	2.7	0.5	0.4	0.5	0.7
FAC2	0.86	0.27	0.72	0.73	0.78	0.56
R	0.57	0.25	0.42	0.36	0.47	0.36

N = ensemble size; obs = observed concentration; mod = modelled value; FB = fractional bias; NMSE = normalized mean square error; FAC2 = fraction of predictions within a factor of two of the observations; R = correlation coefficient.

Table S9. Comparison of 1-h average PM₁₀ concentrations measured in the Sinkule yard with CAMx and PALM results for the same location.
	Summer episodes		Winter episodes		All episodes
	PALM	CAMx	PALM	CAMx	PALM	CAMx
N	130	130	200	197	330	327
mean obs (μg m^-3)	21.1	21.1	28.2	28.4	25.4	25.5
mean mod (μg m^-3)	12.2	8.5	35.8	24.2	26.5	17.9
standard deviation obs (μg m^-3)	7.7	7.7	14.0	14.0	12.4	12.4
standard deviation mod (μg m^-3)	8.1	5.3	20.7	11.9	20.5	12.5
FB	-0.5	-0.9	0.2	-0.2	0.0	-0.3
NMSE	0.6	1.1	0.5	0.3	0.5	0.5
FAC2	0.48	0.29	0.72	0.81	0.62	0.61
R	0.45	0.54	0.32	0.35	0.42	0.46

S3. Infrared measurements and position of evaluation points

Observation location 00-1 (see location IR-01 on Fig. 1 and Tab. S1; IR-00 is same position with extra images for a heat flux validation): the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 01-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 01-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 02-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 02-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Mapy.cz.

Observation location 02-3: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 02-4: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 03-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 03-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 04-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 04-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 04-3: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 04-4: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 05-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 05-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 06-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 06-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation locations 06-3 and 06-4: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 07-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 07-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 08-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 08-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 09-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 09-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 10-1 through 10-3: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 11-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 11-2: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Google.

Observation location 12-1: the view of the observation location and IR and RGB photos with placement of the evaluation points (upper row) and graphs of observed (dots) and modelled (lines) surface temperature for wall (left) and ground (right) for particular evaluation points (EP) for summer e2 (middle) and winter e3 (bottom) episodes. The modelled values come from the child PALM domain, the dotted and dashed lines represent the modelled temperature for left and right grid faces (see Sect. 5.1.1). The grey dashed line shows the corresponding WRF skin layer temperature for horizontal surfaces. The grey areas denote the night time. Top left image © 2020 Mapy.cz.

S4. Modelled surface quantities at IR camera evaluation points

Summer episode 2 campaign

Winter episode 3 campaign

S5. Street canyon quantities

Air temperature

Figure S24. Modelled (solid dark blue line) and observed (dark blue dots) air temperature (T) in particular street canyon observation locations. The modelled values represent the one-hour floating average while the thin dashed black line shows the original 10-minutes averages. The light red band shows the interval between the smallest and the largest value among the neighbouring points of the grid. The thin dotted blue line indicates corresponding value from WRF simulation.

Wind speed

Figure S25. Modelled (solid dark blue line) and observed (dark blue dots) wind speed (WV - wind velocity) in particular street canyon observation locations. The modelled values represent the one-hour floating average while the thin dashed black line shows the original 10-minutes averages. The light red band shows the interval between the smallest and the largest value among the neighbouring points of the grid. The thin dotted blue line indicates corresponding value from WRF simulation.

Surface sensible heat flux

Figure S26. Modelled (solid dark blue line) and observed (dark blue dots) surface sensible heatflux (HF) in particular street canyon observation locations. The observed quantities represent vertical heat flux, corrected for cross wind component neasured by the METEK 3D ultrasonic anemometer at 7 m above ground level. The modelled values represent the one-hour floating average while the darker yellow line shows the original 10-minutes averages. The light yellow band shows the interval between the smallest and the largest value among the neighbouring points of the grid.

Relative humidity

Figure S27. Modelled (solid dark blue line) and observed (dark blue dots) relative humidity (RH) in particular street canyon observation locations. The modelled values represent the one-hour floating average while the thin dashed black line shows the original 10-minutes averages. The light red band shows the interval between the smallest and the largest value among the neighbouring points of the grid. The thin dotted blue line indicates corresponding value from WRF simulation.

Concentration of NO_x

Figure S28. Modelled (solid dark blue line) and observed (dark blue dots) concentrations of NO_x in particular street canyon observation locations. The modelled values represent the one-hour floating average while the thin dashed black line shows the original 10-minutes averages. The light red band shows the interval between the smallest and the largest value among the neighbouring points of the grid. The thin dotted blue line indicates corresponding value from CAMx simulation.

Concentration of PM₁₀

Figure S29. Modelled (solid green line) and observed (green dots) concentration of PM₁₀ in particular street canyon observation locations. The modelled values represent the one-hour floating average while the thin dashed black line shows the original 10-minutes averages. The light red band shows the interval between the smallest and the largest value among the neighbouring points of the grid. The thin dotted blue line indicates corresponding value from CAMx simulation.

Concentration of PM_2.5

Figure S30. Modelled (solid green line) and observed (green dots) concentration of PM_2.5 in particular street canyon observation locations. The modelled values represent the one-hour floating average while the thin dashed black line shows the original 10-minutes averages. The light red band shows the interval between the smallest and the largest value among the neighbouring points of the grid. The thin dotted blue line indicates corresponding value from CAMx simulation.

Mobile observation - concentration of PM₁₀

Figure S31. Mobile PM₁₀ measurements (+ marker, 1 min average) and modelled concentrations (solid, 10 min average) for 19 July 2018 morning (top left), 26 November morning (top right), 4 December 2018 morning (bottom left) and 4 December 2018 afternoon (bottom right). Numbers refer to mobile measurement locations according to Fig. 1

S6. Initial and boundary conditions from WRF

The 3-D fields from WRF outputs (T, Q, U/V/W) were horizontally and vertically interpolated (in that order) to the PALM model grid. Because the PALM model used a higher-resolution terrain that would differ from the coarse terrain in WRF by as much as tens of meters, the vertical interpolation had to include stretching of the atmospheric columns.

At the bottom, the atmospheric columns were shifted to match the PALM terrain, therefore there were no missing data below the original terrain and the surface effects from WRF were preserved. However, at higher altitudes, the atmospheric columns could not be shifted by the same amount, as that would introduce unrealistic horizontal gradients mimicking the terrain shift below. In order to avoid this, the atmospheric columns were stretched heterogeneously. The WRF model uses either sigma or hybrid vertical coordinates, our simulations use the hybrid option where the lowest level is terrain-following and the highest level is isobaric. For each column, the geopotential height of each level in the WRF data was recalculated using the same formula and parameters used in WRF for calculating the heights of the hybrid levels, however with the surface pressure altered to match the PALM terrain. The recalculated level heights were then used for linear vertical interpolation into the PALM Cartesian vertical coordinate system.

The interpolated 3-D fields were used as initial conditions for the first timestep and their top and lateral boundaries were used as boundary conditions for all timesteps. For the velocity fields, the total volumetric flux disbalance was calculated for each timestep as a sum of the volumetric inflow minus outflow for all boundaries. This residual volumetric flux was then divided by the total area of the five boundaries and subtracted from the respective inwards-directed velocity component for each boundary in order to make the inflow and outflow perfectly balanced, as is required by the incompressible equations used in PALM.

The Python code used for processing the WRF and CAMx data into the PALM dynamic driver file has been included into the official PALM distribution and published in the PALM SVN repository since revision 4766 in the directory trunk/UTIL/WRF_interface.