Multilevel analysis of spatial temperature variability in Brno region

Transcription

1 Multilevel analysis of spatial temperature variability in Brno region Petr Dobrovolný, Rudolf Brázdil, Lukáš Krahula Department of Geography, Masaryk University Brno, Czech Republic Czech Science Foundation project no. 205/09/1297 EMS/ECAC, Zurich, 2010

2 Outline 1. Motivation 2. Multilevel approach and UHI as an example 3. Study area and Data available 4. Variability of air temperature at urban and rural stations 5. Spatial differentiation of land surface temperature 6. Summary and outlook

3 Motivation Are we able to detect any signal of higher temperatures within a city compared to its neighborhoods? Urban build-up areas can be considered as a hierarchical system of several levels. Such system is typical with a considerable temporal and spatial differentiation of fields of meteorological elements. Data of different resolution in time and space can be successfully used to analyze spatial and temporal variability of climate in urban environment. The synthesis of different data sources can be valuable for evaluation of potentially increased frequency and intensity of meteorological extremes on the urban environment as a consequence of global climate change. Project no. 205/09/1297 Multilevel analysis of the urban and suburban climate taking medium-sized towns as an example ( )

4 UHI as an example of multilevel approach in urban climatology Boundary layer urban heat island 2. Canopy Layer urban heat island 3. Surface urban heat island (adapted from Oke 1976)

5 Data sources for spatial temperature variability 1. Standard meteorological measurements 2. Special-purpose measurements Position of meteorological stations in Brno (Štěpánek et al. 2006) 3. Methods of remote sensing Air temperature measurements since St. Ann s Hospital 2 Pekařská str. no Augustinian Monastery 4 Technical University 5 Pisárky (waterworks) 6 Květná str. 7 Tuřany (airport) (see poster P4-14 MC2 section)

6 Study area Brno is the second largest city of the Czech Republic (population 400,000, cadastral area 230 sq. km) and is typical by its a basin position with a range of altitudes from 190 m to 479 m. The higher ground lies largely in the western and northern parts of the region, while lower and flatter terrain is typical of its southern and eastern parts. The study area is one of the warmest and, in part, the driest regions of the Czech Republic.

7 Special purpose measurements stations 0 3 km

8 Daily air temperature variability Definition of radiation weather type daily sunshine duration - more than 80% of maximum possible length of duration mean daily wind-speed not higher that 4 m.s -1 smoothed mean curve of daily variability influenced only due to convective clouds Period: 1 Jan Jul days (DJF 4, MAM 19, JJA 35, SON 11)

9 Daily air temperature variability DJF MAM JJA SON

10 Daily air temperature variability DJF MAM JJA SON

11 Difference of mean daily air temperature variability at urban and rural stations In winter urban stations are all day 1 2 C warmer compared to rural stations Occurrence of daily minimum temperature does not differ significantly at urban and rural stations. However, maximum temperature occurs hr earlier at urban stations. Temperature differences between urban and rural stations show a clear daily cycle with maximum at mid-day (more than 2 C in summer). Urban rural temperature differences are minimal in morning (7 8) and evening (17 19 hours).

12 Land Surface Temperatures - LST Digital Numbers DN = f(lst)

13 Study area and data used LANDSAT 7 satellite Scanner ETM+ date 24 May 2001 time 9:35:02 GMT Thermal band μm Spatial resolution 60 m Typical radiation type of weather Synoptic situation NEa T min 8.4 C T max 23.3 C T mean 17.6 C T ground min 5.0 C

14 LST derivation from LANDSAT data ETM+ band 6.1 ETM+ 1,2,3,4 TOA radiance Atmos. corrected values MODTRAN transmissivity upwelling radiance downwelling radiance (Barsi et al. 2005) Emissivity Land cover types Land surface Temperature

15 Emissivity map of basic land cover types water 0,98 bare ground 0,97 vegetation 0,94 built-up area 0,925 Emissivity values (Snyder et al. 1998) This is a weak point of mono window algorithms

16 Land Surface Temperatures, Brno, 24 May 2001 LST

17 Typical LST for different land cover types in Brno region, 24 May 2001 build-up area bare ground vegetation water urban rural

18 UHI Intensity, Brno and surroundings, 24 May NE quadrant 3 4 Distance from the city centre [km]

19 Other RS data - multispectral algorithms TERRA satellite scanner ASTER date 2 April 2002 time 9:57:53 GMT 5 thermal bands (10 14) wavelengths μm spatial resolution 90 m Synoptic situation SEa T min 7.8 C T max 16.8 C T mean 11.3 C T ground min 6.4 C

20 LST derivation from ASTER data ASTER bands TOA radiance Atmos. Corrected values Land Surface Temperatures Emissivity at bands 10-14

21 Summary Different approaches for construction of LST fields have been tested using ETM+ and ASTER thermal imagery LST values in urban areas can reach C. However, LST of areas with a dense vegetation cover vary around 25 C (based on 6 different satellite scenes acquired from April to September) LST of bare ground in southern part of the study area are comparable to those of urban surfaces Land Surface Temperature fields are typical with high spatial variability There is a strong horizontal gradient in LST of N S orientation in Brno area Two different data sources show consistent area of higher temperatures within a city centre

22 Outlook Various parameters derived from 3D model of buildings and from Digital Elevation Model explain spatial variability of land surface temperatures. Area of walls and roofs Area of buildings [%] Impervious surfaces [%]

23 Thank you for your attention

24 Urban climate Selected urban climate variables compared to rural landscape (Landsberg 1981) no. of condensation nuclei no. of solid particles sunshine duration cloud cover frequency of fog in winter precipitation amount frequency of thunderstorms annual mean temperature mean min temperature in winter mean max temperature in summer heating period annual mean relative humidity annual mean wind velocity 10 x more 10 x more 5 15% less 5 10% less 100% more 5 15% less 5 10% more C higher 1 2 C higher 1 2 C lower 10% shorter 6 % lower 20-30% lower Urban Heat Island (UHI)