See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/316055161

Street Canyon Geometry Effects on Microclimate and Comfort; A Case Study

in Thessaloniki

Article  in  Procedia Environmental Sciences · December 2017

DOI: 10.1016/j.proenv.2017.03.144

CITATIONS READS

33 594

2 authors:

Angeliki Chatzidimitriou Kleo Axarli

Aristotle University of Thessaloniki Aristotle University of Thessaloniki
19 PUBLICATIONS   438 CITATIONS    13 PUBLICATIONS   610 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Estimation of heating and cooling loads of residential urban blocks by using urban morphological indicators View project

All content following this page was uploaded by Angeliki Chatzidimitriou on 03 July 2017.

The user has requested enhancement of the downloaded file.

 Available online at www.sciencedirect.com

ScienceDirect
Procedia Environmental Sciences 38 (2017) 643 – 650

International Conference on Sustainable Synergies from Buildings to the Urban Scale, SBE16

Street Canyon Geometry Effects on Microclimate and Comfort; a

case study in Thessaloniki
Angeliki Chatzidimitrioua,*, Kleo Axarlia
a
Aristotle University of Thessaloniki, Faculty of Engineering, School of Architecture, University Campus, Thessaloniki 54124, Greece

Abstract

A case study in Thessaloniki provided microclimate data at street canyons with varying orientation and aspect ratio. Seventeen
canyons characterized by diverse geometries including four different canyon axis orientations and a range of aspect ratios from
0.6 to 3.3 as well as variations of vegetation and pavement material, were monitored during summer and winter. Measurements
are used to calibrate microclimate simulation models, examine results accuracy and produce further data for larger areas or
different climatic periods. This paper presents the simulation studies run with ENVI-met software for models of the study area at
different scale used to evaluate results accuracy depending on model resolution, influence of the surrounding built environment
and of material and geometric details. The large area - low resolution models take into account the urban environment
surrounding the study areas which influences thermal and airflow conditions and provide urban microclimate maps. The small
area - high resolution models take into account fine geometric elements, individual building heights, tree species, canopies, and
pavement materials and provide detailed microclimate data and pedestrian comfort indices which can inform urban space design.
Keywords: Urban microclimate; Outdoor comfort; Case study; Simulation; Software calibration

1. Introduction

Urban morphology is one of the most influential parameters affecting the microclimate of open spaces in cities. In
previous studies the effect of form and materials has been examined in squares, parks, and courtyards through
monitoring campaigns and simulation studies (Chatzidimitriou and Yannas 2015; 2004). Other studies have focused
on the effects of street canyon orientation and aspect ratios (Ali Toudert and Mayer 2006; Emmanuel et al 2007;
Chatzidimitriou and Axarli 2015). However previous research results have not yet been fully interpreted into
guidelines to practically inform urban design and planning.

* Corresponding author. Tel.: +30-697-724-9868; fax: +30-231-083-4371.

E-mail address: ange1iki@arch.auth.gr

1878-0296 © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of SBE16.
doi:10.1016/j.proenv.2017.03.144
644 Angeliki Chatzidimitriou and Kleo Axarli / Procedia Environmental Sciences 38 (2017) 643 – 650

This paper presents a case study in the urban centre of Thessaloniki in Northern Greece, based on a monitoring
campaign at different street canyons in the city centre, followed by a series of simulations of the monitored areas
with microclimate simulation software ENVI-met (.ENVI-met website). The study takes into account summer and
winter measurements but focuses on hot summer conditions and reports on results for ambient temperature,
pavement surface temperature and pedestrian comfort indices of the Physiologically Equivalent Temperature PET
(Hoppe 1999). Simulation models with different size and resolution are tested against measured data to explore the
influence of the models’ details on the results accuracy.
The purpose of the study is to examine the level of agreement between measurements and simulations, as well as
the microclimatic variations developed at the surrounding areas beyond the specified points of the monitored street
canyons. Simulation results by models calibrated with measured data can become tools for extracting climate related
information to aid urban design and planning and to set the basis for new methods, guidelines and new regulations.

Nomenclature

H/W street canyon aspect ratio calculated by building height H to street width W
Tair, Ta air temperature ( oC)
Ts surface temperature (oC)
Tg globe temperature (oC)
Tmrt mean radiant temperature (oC)
PET physiologically equivalent temperature (oC)
WV & WDIR wind velocity (m/s) and wind direction
RH relative humidity (5)

2. Case study

2.1. Location and climate

The city of Thessaloniki in Northern Greece, at 40.5N latitude, is characterized by temperate - mediterranean
climatic conditions, with hot summer periods and cold winter periods, while the vicinity to the coastal zone highly
influences humidity and airflow patterns (sea breeze). The climate parameters during the monitoring campaign in
July 2015 and February 2016, as recorded by the meteorological station at the airport area, are presented in Table 1
and have been used as input for the simulation studies.

Table 1. Climate data during monitoring period (airport meteorological station https://www.wunderground.com/q/locid:16622000;loctype:3).
monitoring period mean Tair (oC) max Tair (oC) mean RH (%) mean WV (m/s) prevailing wind direction
24 – 31 July 2015 26-30 32-38 46-54 1.94-2.78 WNW | S
09-15 February 2016 8-16 12-22 49-92 1.39-3.33 W | NW | SE

2.2. Examined street canyons

Seventeen street canyons characterized by various axis orientation and aspect ratio (building height H to street
width W) within the dense city centre (Fig. 1) were included in the monitoring and simulation studies. In particular,
canyons with four different orientations were examined: seven NW-SE oriented canyons parallel to the coastline, six
NE-SW canyons normal to the coast and four N-S and E-W canyons diagonal to the regular street grid. Mean
canyon aspect ratios varied between 0.6 for the widest canyons to 3.3 for very deep and narrow canyons.

2.3. Monitoring campaign

Monitoring took place in July 2015 (24th-31st) and February 2016 (9th-15th), and included air temperature, relative
humidity, wind velocity and direction and globe temperature measurements at 1.1m above ground level on one side
 Angeliki Chatzidimitriou and Kleo Axarli / Procedia Environmental Sciences 38 (2017) 643 – 650 645

of the canyon, as well as surface temperature readings at both side pavements and at the centre of the streets. The
study was conducted in different groups of cases each day: in canyons with the same orientation and variable street
width and in canyons with different orientation and similar aspect ratio. Measurements were taken in 1 or 2 hour
cycles, with 15 minute intervals between readings at different sites (4-6 sites). The monitoring instrumentation is
listed in Table 2. PET values were calculated from measured data using RayMan (Matzrakis and Mayer 2010), with
Tmrt estimated by measured Tair, Tg and WV using Equation (1) as described by Thorson et al (2007) where D is
the globe’s diameter in mm and ε is the globe’s emissivity.
0.25
ª 1.1 u 10 8 u WV 0.6 º
Tmrt « Tg  273.15  4
 u Tg  Ta ) »  273.15 (1)
¬ HuD 0.4
¼
Measurement results for the summer period are reported in detail in Chatzidimitriou (2016) and reveal the
influence of orientation and aspect ratio on microclimate parameters and pedestrian comfort.

Fig. 1. Locations of the examined canyons in the city centre of Thessaloniki (lat. 40.5 oN).

Table 2. Monitoring instrumentation

variable Sensor | model accuracy resolution
Tair Thermistor | Hobo H08-007-02 (Jul 2015) ±0.7oC 0.33oC
Thermistor | Hobo Pro v2 U23-002 (Feb 2016) ±0.2oC 0.02oC
RH Capacitance | Hobo H08-007-02 (Jul 2015) ±5% 0.1%
Capacitance | Hobo Pro v2 U23-002 (Feb 2016) ±2.5% 0.03%
Ts IR thermometer | Extech 42530 ±0.25oC 0.03oC
Tglobe Thermistor in Grey acrylic sphere D40mm | TM C20 HD (Jul 2015) ±0.25oC 0.03oC
Thermistor in Grey acrylic sphere D40mm | Hobo Prov2 U23-003 (Feb 2016) ±0.2oC 0.02oC
WV Impeller 25mm | Kestrel 4600 ±3% 0.1m/s
WDIR Wind vane & compass | Kestrel 4600 ±5ο 1ο

3. Simulation study

The case study area microclimate in the dense city centre is simulated with ENVI-met v4 BETA . The study area
is modeled in different sizes and resolution to test results accuracy in comparison to measured data. The geometric
models are described below and presented in Figure 2. Simulation input data is listed in Table 3.
x City centre model covers an area 1000mX1000m with 5m horizontal grid resolution and includes all 17 canyons.
x Neighborhood model covers an area 300mX300m with 2m horizontal grid resolution and includes four of the
monitored canyons (sites 01, 02, 03 and 07).
x Single canyon model covers an area 200mX100m with 1m horizontal grid resolution and includes two of the
monitored canyons described in detail (sites 01 and 02).
Vertical grid resolution is 3m in all models, with larger detail in the lower level grid. Building heights vary
between 6m and 31m, with most buildings in the range of 20m - 28m. Sky View Factors deriving from the three
models are shown in Figure 3. Tree heights vary between 3m and 20m and ground level materials in streets and
646 Angeliki Chatzidimitriou and Kleo Axarli / Procedia Environmental Sciences 38 (2017) 643 – 650

pavement profiles are asphalt, concrete tiles, bare soil, low vegetation, and water. Albedo values used for asphalt,
dark concrete pavement, light color concrete pavement and building surfaces are 0.2, 0.2, 0.5 and 0.3 respectively.

a b c

Fig. 2. Simulation model area input files of different size and grid resolution a. city centre model, b. neighborhood model, c. canyon model.

a b c

Fig. 3. Sky View Factors at pedestrian level for the urban areas represented in a. city centre model, b. neighborhood model, c. canyon model.

Table 3. Simulation input data for indicative days

simulation parameters input data
day 26th July 27th July 29th July 30th July
start time, duration 5:00, 24hours 5:00, 24hours 5:00, 24hours 5:00, 24hours
tair range (oC) 21-35 22-37 22-37 23-38
rh range (%) 17-68 18-78 14-69 16.5-65
initial specific humidity (g/kg) 11.3 13.8 12.2 12.4
initial wind velocity (m/s) 2.5 2.78 2.5 2.22
wind direction (degrees) 270 W 293 NW 180 S 293 NW

4. Results and discussion

4.1. Simulation results compared to measured data

38 55
36 Tair measured 50
air temperature (oC)

Tair sim CC
34 45
PET (oC)

32 40

30 35

28 30
PET calculated
25 PET sim CC
26
20
24
site 01 10:00
site 02 10:15
site 03 10:30
site 04 10:45
site 05 11:00
site 06 11:15
site 07 11:30
site 11 11:45
site 08 12:00
site 09 12:15
site 10 12:30
site 13 12:45
site 12 13:15
site 17 13:30
site 14 13:45
site 15 14:00
site 16 14:15
site 01 10:00
site 02 10:15
site 03 10:30
site 04 10:45
site 05 11:00
site 06 11:15
site 07 11:30
site 11 11:45
site 08 12:00
site 09 12:15
site 10 12:30
site 13 12:45
site 12 13:15
site 17 13:30
site 14 13:45
site 15 14:00
site 16 14:15

a b

Fig. 4. Simulation results of the city centre model (CC) compared to measurements for July 30th a. Tair, b. PET

The simulation results for Tair, Ts and PET values are compared to measured data on indicative days. The large
model of the city centre was used to test results from all sites while the smaller size models focus on a specific area
with NW-SE and NE-SW oriented canyons. Figure 4 presents simulation results from the city centre model and
measured values on July 30th. Differences between measurements and simulations range between 0.4 oC and 3.4oC
for Tair and between 0.2oC and 6.0oC for PET and are attributed to the low resolution of the 5mX5m grid which
 Angeliki Chatzidimitriou and Kleo Axarli / Procedia Environmental Sciences 38 (2017) 643 – 650 647

leads to simplifications in geometric forms and pavement materials. Simulation results for pavement surface
temperatures present good agreement with measured data with differences between 0.1 oC and 4.9oC. Large Ts
differences of 9oC, 10oC and 15oC occur only in three cases due to inaccurate shading patterns caused by massive
building heights input. Such differences diminish to 1.5oC by correcting the building heights in higher resolution
models. Ground Ts results within canyons, at both side pavements and the street centre are presented in Table 4.

Table 4. Ground surface temperature simulation results compared to measured data at indicative points within canyons on July 30th
results /site & time Ts measured at Ts sim CC at Ts measured at Ts sim CC at Ts measured at Ts sim CC at the
side pavement side pavement side pavement side pavement the street central street central
NE|SE|E|S NE|SE|E|S SW|NW|W|N SW|NW|W|N axis axis
site 01 10:00 27.2 29.2 27.8 30.0 30.2 29.3
site 02 10:15 29.0 39.1 35.3 35.7 31.8 29.6
site 03 10:30 27.8 37.2 31.4 33.9 40.1 43.6
site 04 10:45 32.0 30.1 40.5 43.6 45.1 43.0
site 05 11:00 32.6 32.8 42.8 44.2 44.5 46.8
site 06 11:15 28.2 29.8 29.6 30.1 28.6 30.1
site 07 11:30 31.0 30.1 29.5 30.7 32.1 30.8
site 11 11:45 28.9 29.7 40.1 42.8 33.5 30.6
site 08 12:00 29.1 29.1 47.2 42.7 38.1 37.2
site 09 12:15 48.4 48.2 45.1 43.3 47.6 49.1
site 10 12:30 28.9 31.6 50.9 51.2 41.3 45.0
site 13 12:45 28.5 31.2 40.1 45.0 38.7 43.4
site 12 13:15 31.5 31.6 40.3 45.1 32.3 31.6
site 17 13:30 32.3 32.2 63.6 48.5 54.7 50.4
site 14 13:45 47.9 46.3 32.5 32.4 31.5 32.4
site 15 14:00 31.5 32.3 39.8 37.8 31.1 32.5
site 16 14:15 32.1 31.0 48.8 46.9 45.5 41.1

4.2. Effects of model size and grid resolution

Examination of different scale simulation models (Fig. 5), reveals higher accuracy with increasing grid
resolution; however in most cases the best agreement with measured values is observed with the neighborhood
model (medium size). In comparison to the city centre model results (large scale and low resolution), the average
differences between measurements and simulated values (Table 5) reduce
x for DTair from 1.7oC (city centre model) to 1.0oC (neighborhood model),
x for DTs from 0.8oC (city centre model) to 0.4oC (neighborhood model) and
x for DPET from 2.0oC (city centre model) to 0.2oC (neighborhood model) and to 0.1 oC (single canyon model).

a b c

Fig. 5. Simulation results for PET at 15:00 on July 26th a. city centre model, b. neighborhood model, c. single canyon model.

Considering distinct monitoring cycles, the highest differences between measurements and simulations, with all
three models, are observed around noon for air temperature (11:00-12:30 and 13:00-14:30) and in the morning and
afternoon cycles for PET (9:00-10:30 and 15:00-16:30). A more detailed comparison between measured and
simulated values is presented indicatively in Figure 6 for morning and afternoon hours.
The results show the best agreement with medium size and resolution model which contains several building
blocks (neighborhood scale model), indicating that relatively high resolution is required to achieve adequate
geometric detail but also it is significant to include a considerable part of the surroundings to obtain satisfactory
results for the area of interest. The discrepancies observed in the large city centre model are not very high and they
648 Angeliki Chatzidimitriou and Kleo Axarli / Procedia Environmental Sciences 38 (2017) 643 – 650

are attributed to reduced geometric detail and bulk representation of building heights which alter incident solar
radiation on surfaces. The differences observed in the single canyon model are in many cases similar to those in the
neighborhood model and higher divergence may be partly due to the absence of sufficient buildings around the
canyons which influence airflow and also due to urban equipment and miscellaneous construction elements which
are not included in the geometric model. The presence of additional construction elements at pedestrian level and on
building surfaces as well as the detailed representation of individual trees are issues that need to be considered in the
development of small scale models with high resolution in order to increase accuracy.

Table 5. Mean difference of simulation results by three different models (city centre CC, neighborhood NH and single canyon model SC)
compared to measured data for July 26th. DT refers to the difference resulting by subtracting the simulation results from measured values.
measurement cycle DTair sim DTair sim DTair sim DTs sim DTs sim DTs sim DPET sim DPET sim DPET sim
(time range) CC (oC) NH (oC) SC (oC) CC (oC) NH (oC) SC (oC) CC (oC) NH (oC) SC (oC)
9:00-10:30 1.0 0.3 2.0 -1.6 0.5 1.0 -3.1 2.2 3.9
11:00- 12:30 3.4 2.7 3.5 -0.8 -0.8 -0.5 -1.5 0.0 2.2
13:00-14:30 3.0 2.0 1.8 1.1 2.4 -0.6 -0.7 -0.4 -1.4
15:00-16:30 0.8 0.1 -0.3 -1.5 -1.0 -1.4 -3.4 -3.1 -3.7
17:00-18:30 0.4 0.0 -1.3 -1.0 1.0 -1.5 -1.3 0.2 -1.6
daily average 1.7 1.0 1.1 -0.8 0.4 -0.6 -2.0 -0.2 -0.1

34 34
Tair measured
32 Tair sim CC 32
air temperature (oC)
air temperature (oC)

Tair sim NH
Tair sim SC
30 30

28 28 Tair measured
Tair sim CC
Tair sim NH
26 26
Tair sim SC
24 24
a b
site 01 site 02 site 03 site 04 site 05 site 06 site 07 site 01 site 02 site 03 site 04 site 05 site 06 site 07
9:00 9:15 9:30 9:45 10:00 10.15 10.30 15:00 15:15 15:30 15:45 16:00 16.15 16.30
42 42
40 Ts measured 40
sufrace temperature (oC)
sufrace temperature (oC)

38 Ts sim CC 38
Ts sim NH
36 36
Ts sim SC
34 34
32 32
30 30 Ts measured
28 28 Ts sim CC
26 26 Ts sim NH
24 24 Ts sim SC
22 c 22 d
site 01 site 02 site 03 site 04 site 05 site 06 site 07 site 01 site 02 site 03 site 04 site 05 site 06 site 07
9:00 9:15 9:30 9:45 10:00 10.15 10.30 15:00 15:15 15:30 15:45 16:00 16.15 16.30

42 42
40 PET calculated
40
PET sim CC
38 PET sim NH 38
36 PET sim SC 36
PET (oC)

PET (oC)

34 34
32 32
30 30 PET calculated
28 28 PET sim CC
26 26 PET sim NH
24 24 PET sim SC
22 22
e f
site 01 site 02 site 03 site 04 site 05 site 06 site 07 site 01 site 02 site 03 site 04 site 05 site 06 site 07
9:00 9:15 9:30 9:45 10:00 10.15 10.30 15:00 15:15 15:30 15:45 16:00 16.15 16.30
Fig. 6. Simulation results with three models of different size and resolution (city centre scale CC, neighborhood scale NH and single canyon scale
SC) compared to measured data for July 26th in the morning and the afternoon; a. and b. air temperature; c. and d. pavement surface temperature;
e. and f. PET in the morning; and in the afternoon respectively

4.3. Canyons of different orientation

Simulation results by the large city centre model have been used to examine the maximum differences between
canyons with varying orientation and similar aspect ratio. Four canyons are compared, with mean aspect ratio
 Angeliki Chatzidimitriou and Kleo Axarli / Procedia Environmental Sciences 38 (2017) 643 – 650 649

approximately 1.7 and axis orientation perpendicular to the coast NE-SW, parallel to the coast NW-SE, and diagonal
to the regular urban grid, E-W and N-S. The canyons were included in the monitoring campaign as sites 08, 09, 10
and 11 respectively. Simulation results of PET and air temperature are examined for the 27 th of July.

Fig. 7. PET Simulation results at 1.5m above ground level on July 27th in the afternoon (15:00)

Regarding pedestrian comfort indices PET, (Fig. 7), the most favorable conditions at summer afternoon are
observed on the west side of the N-S oriented canyon, site 11, with 37.0oC PET at 15:00. Although this value
indicates strong heat stress it is the lowest observed among the four examined canyons at that time. The N-S canyon
is the most comfortable also in the morning with 26.9oC PET at 9:00. The most unfavorable site in the afternoon,
requiring improvement, is the E-W oriented canyon (site 10) with 48.7oC PET at 15:00 at the north side. This high
PET value indicates extreme thermal stress for pedestrians and is attributed to direct solar radiation, pointing the
need for additional shading, while the results under trees show lower PET by 5oC-10oC at the north side and by 11oC
at the south side. In the morning most of the north side of the E-W canyon is comfortable with PET values around
27oC at 9:00 but high PET up to 42oC occur at individual insolated spots. The measurements taken on site, show
high globe and surface temperatures on the north side all day (10:00-17:00) because of solar exposure, but quite low
surface temperature under the trees. In the morning hours, high PET values, up to 48oC at 9:00, are also observed at
the SW-NE canyon (site 09) due to temporary insolation. At night PET differences among differently oriented
canyons are negligible, below 1oC; simulation results at 23:00 show values between 25oC and 25.6oC at the 4 sites.
Similar effects are observed in deeper canyons as well, at sites 12 and 13, with aspect ratio around 3, oriented E-
W and N-S respectively. The simulation results at 15:00 in the afternoon show 54.8 oC PET in the north side of the
E-W canyon and 37.5oC in the west side of the N-S canyon, similarly to the medium width canyons.
Regarding air temperature, simulation results in the four canyons show negligible differences, below 0.5 oC, all
day. Tair ranges between 27.3 oC and 27.5oC in the morning, 29.9oC and 30.1oC at midday, 32.5oC and 32.8oC in the
afternoon and 29.8oC and 30.1oC at night. Similar conclusions are drawn by on site measurements; Tair does not
vary significantly among sites indicating that axis orientation, at moderately deep canyons, mostly influences PET.

4.4. Canyons of different aspect ratio

The city centre simulation model results are also used to check the effects of aspect ratio variations, in canyons
with the same axis orientation. Three NE-SW oriented canyons, monitored as sites 15, 16 and 17, with average
aspect ratios of 3.3, 2.8 and 1.1 respectively, are examined with regards to PET and Tair on July 29th. Regarding
pedestrian comfort indices PET, the most favorable canyons at midday and afternoon are the deep canyons at sites
15 and 16, with PET values around 35oC at 12:00 and 38oC at 15:00 at both side pavements. However, in the
afternoon even at the narrowest canyon, there are insolated spots, near the crossroads, which present conditions of
extreme heat stress with PET values up to 58.6oC. Unfavorable conditions requiring improvement are observed at
the wide canyon at site 17; PET values reach up to 49.6 oC in the NW insolated side at noon and up to 55.2oC at the
SE insolated side in the afternoon. However at the shaded sides (SE at noon and NW in the afternoon) conditions are
similar to the narrow canyons, in terms of comfort. The simulation results also show lower PET under trees; up to
11oC lower at the NW side at noon and up to 13 oC lower at the SE in the afternoon. At night the differences are
negligible and PET ranges from 24.4 oC to 24.8oC at 2300 in all three examined sites.
650 Angeliki Chatzidimitriou and Kleo Axarli / Procedia Environmental Sciences 38 (2017) 643 – 650

Air temperature simulation results at the NE-SW canyons are similar regardless of aspect ratio. Tair values on
July 29th fluctuate between 26.8oC and 27.5oC in the morning, between 30.3oC and 31.1oC at noon, around 33oC in
the afternoon and between 29.2oC and 29.8oC at night, with the slightly lower values at the narrow canyons.

5. Conclusions

A monitoring study in street canyons in Thessaloniki has been used to calibrate and evaluate microclimate
simulation models with ENVI-met software for a comparative analysis of the effects of geometric characteristics on
microclimate and pedestrian comfort. The comparison between measurement and simulation results of air
temperature, surface temperature and comfort indices PET, revealed good agreement with three different size
models and the best agreement with the medium size model. The moderate grid resolution and representation of the
surrounding environment in the medium model appears adequate and provides the most accurate results compared
to measured data. Nevertheless the large size model offers a view of the microclimate simultaneously in diverse
canyons and open spaces providing maps that may lack in resolution but allow comparisons among various sites.
Small scale model proved sufficient in terms of accuracy, however, high grid resolution requires highly detailed
geometric input to be efficient and should be used for evaluating specific interventions explicitly described. In
general, the overall efficiency of both the large and small scale models largely depends on the scope of the analysis.
The simulation models were used for a preliminary comparative analysis of the effects of aspect ratio and canyon
orientation. The results indicated that the most comfortable conditions in summer occur in N-S oriented canyons of
medium or high aspect ratio while E-W oriented canyons need improvements such as additional shading on the
exposed north side. Deep canyons also proved more comfortable than wide ones in summer afternoon while no
differences were observed at night. Further analysis of both measured and simulated data of summer and winter
conditions will provide more insights on the effects of urban morphology on the microclimate and potentially tools
for suggesting design guidelines and regulations for improving the urban environment, enhancing outdoor comfort
and reducing energy consumption requirements.

Acknowledgements

This study is an ongoing research funded by “IKY fellowships of excellence for postgraduate studies in Greece –
Siemens program”.

References

1. Chatzidimitriou, A and S. Yannas (2015). Microclimate development in open urban spaces: The influence of form and materials. Energy and
Buildings 108 (2015) 156-174. http://dx.doi.org/10.1016/j.enbuild.2015.08.048
2. Chatzidimitriou, A. and S. Yannas (2004). Microclimatic Studies of Urban Open Spaces in Northern Greece. In proc. of 21st PLEA
conference, Built environments and environmental buildings. Eindhoven, The Netherlands.
3. Ali-Toudert, F. and H. Mayer (2006). Numerical study on the effects of aspect ratio and orientation of an urban street canyon on outdoor
thermal comfort in hot and dry climate. Building and Environment Vol. 41, pp94-108.
4. Emmanuel, R., H. Rosenlund and E. Johansson (2007). Urban shading - a design option for the tropics? A study in Colombo, Sri Lanka.
International Journal of Climatology vol. 27, pp1995-2004
5. Chatzidimitriou, A. and K. Axarli (2015). Microclimate development and comfort in urban canyons in relation to orientation and aspect ratio;
exploring guidelines for urban design regulations. In proc. of 31st PLEA Conference, Architecture in (R)evolution., Bologna, Italy.
6. ENVI-met model website : http://www.model.envi-met.com/hg2e/doku.php
7. Hoppe, P. (1999). The Physiologiacal equivalent temperature - a universal indexfor the biometeorological assessment of the thermal
environment. International Journal of Biometeorology Vol. 43, pp71-75
8. Matzarakis, A., F. Rutz and H. Mayer (2010).Modelling radiation fluxes in simple and complex environments: basics of the RayMan model.
Int J Biometeorol (2010) 54:131–139 DOI 10.1007/s00484-009-0261-0
9. Thorsson, S., F. Lindberg, I. Eliasson and B. Holmer (2007). Different methods for estimating the mean radiant temperature in an outdoor
urban setting. International Journal of Climatology Vol.27 pp1983-1993.
10. Chatzidimitriou, A. (2016). Street Canyon Geometry and Microclimate: A monitoring study in Northern Greece. In proceedings of the 32nd
PLEA Conference: Cities, Buildings, People: Towards Regenerative Environments, Los Angeles, USA.

View publication stats