Modelling the Atmospheric Urban Heat Island and its Contributing Spatial Characteristics

Transcription

1 the Atmospheric Urban Heat Island and its Contributing Characteristics The Case of The Hague, the Netherlands Anna-Maria Ntarladima 07/11/2016

2 Urban Heat Island (UHI) is a phenomenon in which a city forms microclimates which heatup quicker than its rural surroundings. ΔΤ(u-r) = Tu- Tr (Oke,1982;, Johnson & Wilson,2009; Arnfield, 2003; Taha et al.,1997; van der Hoeven & Wandl, 2015; Li et al., 2011; Steeneveld et al., 2011) 2

3 There are different UHI types depending on the layer in which the temperature is measured UBL Surface UHI Atmospheric UHI Upper Atmospheric UHI LST UCL (EPA, 2008) 3

4 The UHI types share different characteristics in terms of temporal development, intensity and spatial distribution LST UBL UCL (EPA, 2008) 4

5 The UHI was assumed to have minor effects in The Netherlands till the first major heat-wave (2003) UBL Unhealthiness Mortality Discomfort Respiratory disorders (Garssen, 2005; Hoeven & Wandl, 2015; Mavrogianni, 2011) 5

6 The aim of this thesis is to model the AUHI The scope of this research is to bridge the gap between the raw data and the qualitative research capabilities of planners and designers enable them design more livable cities How does the AUHI develop timely and spatially? How can the air temperature be modeled in a dynamic way? 6

7 UBL The Municipality of The Hague The Area of Interest 7

8 UBL Netatmo weather station Netatmo weather map 8

9 UBL Netatmo weather station Netatmo sensors utilized 9

10 Q* = QE + QH + QS Datasets derived: QE QH Vegetation Index Water Surfaces Imperviousness Albedo Distance from the Coastline QS Building Footprint Height of the buildings Building Volume Sky View Factor 10

11 Datasets used: Sensor data Top10NL AHN3 Imperviousness (Copernicus) Satellite images KNMI 330 and 344 temperatures CBS grid Datasets derived: Land Surface Temperature Vegetation Index Water Surfaces Albedo Distance from the Coastline Building Footprint Height of the buildings Building Volume Sky View Factor Imperviousness 11

12 A certain methodology was applied in ArcGIS All the datasets were converted in the same format and All data were organized in the same grid In all grid-cells a value of each spatial variable is assigned Satellite image, Landsat 8, Band 4 12

13 A certain methodology was applied in ArcGIS All the datasets were converted in the same format and All data were organized in the same grid In all grid-cells a value of each spatial variable is assigned Vegetation Index raster format 13

14 A certain methodology was applied in ArcGIS All the datasets were converted in the same format and All data were organized in the same grid In all grid-cells a value of each spatial variable is assigned 100m * 100m CBS grid combined with a fishnet 14

15 A certain methodology was applied in ArcGIS All the datasets were converted in the same format and All data were organized in the same grid In all grid-cells a value of each spatial variable is assigned Vegetation Index vector format 15

16 Organize all data into the grid Example: Footprints, Values: Example: Coastline Proximity, Values: 0 max. distance UBL Footprints 16

17 Organize all data into the grid Example: Footprints, Values: Example: Coastline Proximity, Values: 0 max. distance UBL Coastline Proximity 17

18 LST NDVI Water Surfaces Albedo Coastal Proximity B. Heights Footprints SVF Imperviousness B. Volume 18

19 19

20 The spatial variables have different behavior during day and night Thought, the day was divided in day (05:27) and night (22:06). Air temperature, has smaller heat capacity temperature changes quicker Day division based on the inflection points 7 datasets derived UBL Temperature-time graph 20

21 Adjusted Std. Error of the R Square Change Change Statistics F Change df1 df2 Model R R Square R Square Estimate Night.639 a a. Predictors: (Constant), SVF, KNMI_344, Coastline Proximity, Albedo, Footprint b. Dependent Variable: aver_s_tem Adjusted Std. Error of the R Square Change Change Statistics F Change df1 df2 Model R R Square R Square Estimate 05_ a a. Predictors: (Constant), Imperviousness Copern, KNMI_344, Albedo, NDVI b. Dependent Variable: aver_s_tem 21

22 22

23 Cooling-down model (22:00 06:00) 23

24 Heating-up model (06:00 13:00) 24

25 Dynamic Time is used as input data Animated maps enable large data-series visualization More perceivable Easy pattern and trend detection If animated visualization was not utilized in this study, 80 different maps would have been presented instead of concise videos 25

26 26

27 27

28 Surface temperature and atmospheric temperature behave differently Ts has strong and significant relation with most indicators Ts is not related with the Ta Ta, due to the small heat capacity and air movements, is more prone to temporal changes greater day division is needed Correlations Day Night aver_s_tem LST aver_s_tem LST aver_s_tem KNMI_ ** ** KNMI_ ** LST NDVI ** ** Albedo ** **.244 ** Water ** ** Coastline Proximity.154 ** **.166 ** ** Footprint *.460 ** *.452 ** Buildings Height **.171 ** *.177 ** Building Density **.211 ** *.214 ** SVF.063 **.173 ** ** Imperviousness ** ** **. Correlation is significant at the 0.01 level (1-tailed). *. Correlation is significant at the 0.05 level (1-tailed). 28

29 The day and night models are affected from different indicators The diurnal model is affected by NDVI, Albedo, Imperviousness The nocturnal model by albedo, Coastline proximity, Buildings Both have strong relation with the rural Ta 29

30 Limitations Uncertainty of the sensor data quality Sensor orientation leads to different results Biased sensor sample Ta is highly affected by air movements and circulation Recommendations Sensor installation Use more sensors Acquire information about the sensor orientation Model the air movements (CFD) 30

31 Adaption Use lighter colors (increase albedo) Use less impervious materials Increase Vegetation Sensor installation 31

32 Adaption Street orientation (vertical to the coastline) to maximize the see breeze effect Lower building coefficient close to sea. The complex urban environment generates more turbulences and wind shear due to high-rise buildings More research is needed to simulate the air movements 32

33 Adaption Street orientation (vertical to the coastline) to maximize the sea breeze effect Lower building coefficient close to sea.? The complex urban environment generates more turbulences and wind shear due to high-rise buildings More research is needed to simulate the air movements 33

34 Thank you for the attention

35 Questions