Remote Sensing of Urban Surface Temperatures

Transcription

1 Remote Sensing of Urban Surface Temperatures James Voogt University of Western Ontario London, Ontario, Canada This work is supported by the Natural Sciences and Engineering Research Council of Canada. 1

2 2 Urban Surfaces: Exhibit strongly threedimensional surface structure M. Roth NUS Show microscale variation of surface properties J. Voogt UWO This is obvious when the spatial resolution is high.

3 Surface Urban Heat Island Day Day Large positive values both day and night Strong control by surface properties. Hot hot Night cool Cold N 0 15 km Traverse 0 C 15 C Surface and atmosphere linked via the energy balance. Look at the details within a pixel: surface types and structure 3

4 Thermal Remote Sensing of Urban Surfaces How does the sensor view the surface? FOV atmospheric effects Sensor characteristics (FOV, viewing geometry) 3-D urban surface representation Solar geometry: sunlit/shaded surfaces off-nadir angle N εσt 4 ε(1 L ) Azimuth Radiometric Source Area (Soux et al. 2004) 4

5 Urban effective thermal anisotropy early afternoon Three-dimensionally rough urban surfaces create effective thermal anisotropy (angular variation in remotely-sensed temperature), visualized on a polar plot: N Hot Cold Sensor view direction sensor view angle 5

6 6 Anisotropy and the SUHI How does urban thermal effective anisotropy impact our interpretation of the SUHI? o anisotropy means that remotely-determined surface temperature values are potentially dependent on viewing conditions sensor viewing position, surface structure, and time of day o therefore SUHI values will be impacted by those same considerations What is known about urban thermal anisotropy?

7 Observed Urban Thermal Anisotropy Marseille City Centre Morning (Lagourade et al. 2004) Surface Anisotropy ( C) Grass, Crops Row crops Forests 2 7 Mountainous Terrain 3 10 Urban Areas 4 10 is large with respect to natural surfaces is complicated and expensive to undertake may be spatially and/or temporally limited may be difficult to generalize from: specific meteorological conditions microscale surface variability specific morphology of the city Vancouver: Light Industrial Mid-morning, 45 ONA View Voogt and Oke (1997) 7

8 8 Extending Observations: Use of Models Observed facet surface temperatures TUF-3D (Krayenhoff & Voogt 2007) + Sensor view model and surface GIS SUM (Soux et al. 2004) T rad for sensor view direction Vancouver Downtown late morning 31 different locations Standard deviation

9 Model Application: Origins of Anisotropy Modelled Temperature ( 蚓 ) anisotropy Observed GIS surface Fixed surface Observed (o) Modelled: GIS surface (m) Modelled: Fixed surface (f) No So Eo Wo Nm Sm Em Wm Nf Sf Ef Wf Voogt (2008) View Direction and Simulation the surface structure only accounts for about 50% of the observed anisotropy incorporation of microscale temperature variability is important (at the land-use scale) 9

10 Extending Observations: Coupled Models Urban surface energy balance model Hourly visualization of TUF-3D & SUM modelled remotely-sensed brightness temperature August 15, 1992 TUF-3D (Krayenhoff & Voogt 2007) + Sensor view model and surface GIS SUM (Soux et al. 2004) Vancouver LI: modelled angular variation of brightness temperature local solar time (hourly) s = sun position 10

11 11 Summary Urban effective thermal anisotropy (UETA) o is large both relative to other surface types and to other factors in the remote calculation of urban surface temperature (atmospheric correction and surface emissivity) o is traditionally not incorporated in SUHI analysis o is dependent on both surface structure and variations in material properties.

12 12 Anisotropy and the SUHI Can we take into account the urban thermal anisotropy when observing the SUHI?

13 Urban Surfaces: Conceptual Definitions Complete Ground-level Roof-top Bird s-eye or plan view Screen-level Zero-plane displacement Voogt and Oke (1997) 13

14 1. A complete surface temperature? 5000 Area (m 2 ) N 100 m Complete surface area nadir (plan view) temperature Incorporate building wall and vegetation temperatures from vehicle traverses 42 C Surface Temperature Complete surface temperature Voogt (2000) 14

15 Correction of SUHI using Complete Surface Temperature Tair Tsfc (3-D corr) Tsfc (nadir) Temperature ( 蚓 ) Agricultural Mixed Rural Warehouse Residential Night: 0530 LDT; Sunset + 9 hrs Day: 1530 LDT; Solar Noon + 2 hrs Res./Comm. Lt. Indust. Downtown Park water (tunnel) water (bridge) water (bridge) Traverse Distance (km) Voogt and Oke (2003) 15

16 Corrections and SUHI Interpretation Temperature Correction ( 蚓 ) ? Day T ac -T b : atmospheric correction T dr -T ac : emissivity correction T 3d -T dr : 3-D surface correction? Night T ac -T b : atmospheric correction T dr -T ac : emissivity correction T 3d -T dr : 3-D surface correction Night Agric Warehouse Comm/Res Downtown Mixed Rural Residential Lt. Indust. T ac -T b atmospheric correction T dr -T ac emissivity correction T 3d -T dr 3-D surface correction Agric Warehouse Comm/Res Downtown Mixed Rural Residential Lt. Indust. UHI Day UHI Night Surface UHI ( 蚓 ) SUHI Tbright Tatmcorr Tdir rad T3-D Tair 16

17 17 Satellite Time Series directional corrections are significant and should be systematically done for temporal analysis of AVHRR data time series (Lagouarde et al. 2004) Corrections of up to ±4 C may be required

18 18 2. A slightly different approach K L u 33 m T air (canyon) T air (rural) u canyon Heat Island Magnitude ( C) morning: cool walls nightime: cool roofs, warm walls SUHI 3-D SUHI 2-D CLUHI pm 12 am Time (Decimal YD) Ground-based wide FOV sensor that incorporates a view of all surfaces (allows for calculation of SUHI 3-D) urban: Basel Sperrstrasse (Dense urban) rural: Basel - Lange Erlen (Grassland)

19 19 Summary Incorporation of 3-D surface structure can change interpretation of SUHI. o Daytime: reduces SUHI (from bird s eye view) cool walls o Nighttime: increase of SUHI warm walls No commonly applied method for including UETA impact on SUHI. Models provide promise in evaluating general representation of UETA that may be useful for applications (e.g. impact on SUHI interpretation)