The analysis was divided based on the research questions generated in achieving the objective of understanding the effects of various built forms of Port City, Colombo on pollution dispersion within itself and to existing Colombo.
4.1 Morphology and pollutant dispersion within Port City and existing Colombo
The following analysis is discussed under 2 sections;
- Does the morphology of the context and peripheral buildings of the study area impact pollutant dispersion and wind paths?
- Does the proposed built forms and pedestrian networks affect pollutant dispersion and wind paths?
4.1.1 Morphology of context and peripheral buildings
Buccolieri et al (2010) found that the breathability of the urban canopy layer is strongly dependent on the building packing density. At higher densities the city responds to wind as a single obstacle while at lower densities it responds as an agglomeration of obstacles. Most urban cities would have a street network or skimming flow where the spacing between buildings is much less than the building height (Shen et al. 2015).
As shown in Figure 03; due to the building height and proximity between buildings, a skimming flow was seen over the study area in all simulations. Variations in vortices and flow patterns were formed within canyons of different built forms. Two distinct wind paths were formed on either side of the study area due to the main streets moving from west to east. As seen by the CO mass fraction contour maps in Table 02, due to the differences in form and orientation of buildings along the periphery of these streets, pollutant dispersion patterns changed with each built form.
Buildings oriented along the wind direction at the northern boundary of the study area directed pollution generated at the street towards the existing context downwind (Scenarios 1,3 & 4). In Scenario 02, the buildings were oriented perpendicular to the wind direction and therefore stagnate pollution within the streets itself and reduce dispersion towards existing Colombo.
Similarly, at the southern boundary, building orientations of scenario 02 & 04 obstruct the ambient flow along the street and redirect wind inwards to the study area. In Scenario 02, this results in pollution being stagnated within the street network while in scenario 04 it is pushed towards existing Colombo. Streamlines of ambient wind from inlet 01 were seen to move through the buildings of the southern boundary in scenario 03 creating areas with low CO mass fractions.
4.1.2 Varying morphology of buildings within the study area
As shown by the CO mass fraction contour maps of Table 03, distinct differences were seen in pollutant dispersion patterns when comparing between models.
The built form of Scenario 03 allowed the maximum amount of ambient wind to flow inland. In this case, pollution generated at street level was pushed west resulting in lower concentrations within the study area, but higher concentrations at the buffer zone and existing Colombo situated downwind of the study area.
In comparison, Scenario 02 offers the least ambient wind movement and therefore least concentration of pollution moving to the immediate context. However, this resulted in the pollution inlet or the vehicle source to be the dominant velocity and as detailed in the following sections, the pollution plume moved upwards along the canyon and dispersed at higher elevations compared to other built forms.
In scenario 04, while maintaining the same ground floor area, the number of buildings were increased thereby increasing the number of spaces between buildings and the overall porosity of the built fabric. This created a more dispersed wind flow within the study area. Discontinuous streets within inner grids and dense tall buildings at the perimeter were seen to worsen ventilation conditions (Maing 2022). This same scenario is seen when comparing scenarios 02 and 04. Super-blocks as created by the layout of scenario 02 need to be more porous to encourage ventilation. The inner voids of scenario 04 (contour maps of table 03) with more porous buildings at the periphery resulted in better pedestrian spaces with lower CO mass fraction readings. However, the increased porosity at the periphery and consequent dispersed wind flow resulted in scenario 04 having the widest spread pollutant plume moving towards existing Colombo.
4.2 Effect of Podiums on pollution dispersion
The addition of a block or tiered podium reduced the spacing between buildings, thereby reducing possible wind paths at street level. However, while usable plot area increased by as much as four times in some cases, CO mass fraction and velocity readings at street level showed minimal deviation (Table 07).
The block podium concentrated street pollution within the podium height with a vortex forming at the foot of the podium. The downward trajectory of the wind along the face of the tower caused an area with good air quality at the top of the block podium (Figure 4). A re-circulation zone formed at the foot of the tiered podium was seen to push the pollution upwards along the face of the podium and tower. This resulted in lower wind velocity and pollutant concentrations at street level (Figure 4).
At higher levels such as at 50m and 100m heights, the effects of the podium were reduced, but not negligent. The pollutants blocked at street level were pushed upwards and dispersed at upper levels. Higher levels of turbulent flows were also seen in models with block podiums in comparison to tiered podiums.
4.3 Effect of canyon geometry on pollution dispersion
A skimming ambient flow was seen over the cluster due to its height and proximity (Figure 3 & 4). This supports existing knowledge of flow patterns (Zajic et al. 2011). Probe readings of CO mass fraction show that all pollution was dispersed once it reached the ambient skimming flow, until such time it showed fluctuating readings within the canyon. CO mass fraction readings reached zero while velocity readings reached a uniform value similar to the ambient wind speed. The height at which the skimming flow was reached is dependent on the canyon geometry (Figure 5).
The dominant factor affecting pollution dispersion was the height of the windward tower in comparison to the leeward tower. At probe A, the skimming flow was formed at the top of the shorter windward building, after which the pollutant concentration reduced rapidly (Figures 04 & 05). A shorter leeward tower and a taller windward tower (at probe F) were seen to create an elongated vortex (Figure 04), pushing the pollutant plume over the downwind building. Equal building heights and a narrow canyon at probe B, resulted in a skimming flow over the top of the towers. A series of independent vortices were formed within the canyon due to its narrow width. Table 04 shows that with each building form, the height, width, and street geometry (podium) of each canyon changed. This resulted in significant differences in dispersion patterns seen within each canyon.
While velocity readings showed a similar pattern of reaching ambient speeds over the urban canopy layer, it does not follow the same fluctuations of the CO mass fraction values within the canyon. This suggests a low correlation between the two parameters.
The building cluster studied here has both uniform (Probe B), non-uniform canyons (Probe A & F) and isolated building geometries (Probe G). Readings of CO mass fraction at street level showed high values for B & F for all models, with varying values for A & G. Studies suggest uneven building layouts help improve the dispersion of urban pollutants as the maximum and mean concentrations recorded in simulated non-uniform street canyons at pedestrian levels were found to be less than those of uniform canyons (Gu et al, 2011). While readings from Probe A & B support this, the street level readings from probe F do not.
CO mass fraction and velocity readings at street level at probe F showed minimal variation between built forms and overall recorded high values compared to other probes. B is a uniform deep canyon while F is a non-uniform deep canyon. This suggests that within a deep, narrow canyon such as at Probe F with reduced ambient wind flow, the impact of varying building heights on street level dispersion is low. However, when comparing vertical CO mass fraction readings, higher fluctuations were seen at probe F than at probe B. Beyond 150m height, CO mass fraction readings of Probe B followed a similar path while at F significant variations were seen with changing morphology. Scenario 02 (T02, TP02 & TTP02) showed pronounced effects due to the further lack of ambient wind flow.
4.4 Effectiveness of urban morphological parameters as regulatory guidelines
The morphological parameters of plot coverage, FAR, building separations, frontal area density and permeability were correlated with wind velocity and CO mass fraction readings at street level. Data from the study showed significant differences in CO mass fraction and velocity readings between models even when the same plot coverage and floor area were used.
4.4.1 Building Separations
As mentioned in section 4.1, each building within the cluster responded in a unique way to the ambient wind and pollution generated due to its location within the cluster. Depending on its location and orientation to the dominant wind path, the standard building separations either blocked or assisted pollution dispersion. Studies have shown that building porosity can be increased by adding a larger distance between buildings resulting in better wind movement (Yuan et al., 2014). However, as seen in the study, this does not result in more effective pollutant dispersion as the path taken by the pollution varied with each built form and had a higher impact on readings at street level.
In all simulations the total frontal facade length remains constant. However, due to variations in number of buildings and their facade length, the average spacing between buildings vary. As such, as shown in Table 05; Case 01 (towers only), showed the highest average building separations between towers at ground level. Of this, T01 has the largest while T04 has the smallest average building separations.
A low correlation was seen between building separations of Case 01 and CO mass fraction at street level. However, a strong negative correlation was seen in Cases 02 & 03. The addition of a podium reduced the spacing between buildings significantly, which had a higher impact on pollution dispersion at ground level than when there was no podium.
Model T04 has the shortest building separations and showed high CO mass fraction readings at street level. However, CO mass fraction contour maps (Table 02) showed pockets of low pollution concentration within each block along the pedestrian walkways (North-south and East-west access routes within each block). In this instance, the smaller building separations reduced dispersion within streets, but they also assisted in sheltering internal pedestrian networks (Table 02) from being exposed to pollution generated within the street.
4.4.2 Frontal Area Density and Permeability
The following correlation relationships were determined from the data extracted.
- Frontal area density and permeability consider the front façade of buildings facing the windward direction. A strong negative correlation was seen between average permeability and average frontal area density (Table 07).
- In all cases (tower only, tower + block podium, tower + tiered podium), CO mass fraction at street level showed a strong positive correlation with average permeability. This means street level pollution concentration was heavily impacted by the permeability capacity of building facades.
- Velocity readings at street level showed mixed correlation with permeability in each model. A strong positive correlation with average permeability was seen in case 02 (tower + block podium), a strong negative correlation in case 03 (tower + tiered podium) and a weak negative correlation in case 01 (tower only).
Table 7. Comparison of average frontal area density and average permeability against CO mass fraction and velocity readings at street level of 1.5m
Case 1
|
Average Permeability
|
Frontal Area Density
|
Average Building Separation
|
Average readings at street level of 1.5m
|
Tower only
|
|
|
|
CO mass fraction
|
Velocity (m/s)
|
T01
|
0.5009
|
0.1628
|
80.06
|
0.74
|
1.85
|
T02
|
0.3037
|
0.22641
|
49.78
|
0.62
|
1.73
|
T03
|
0.6719
|
0.10645
|
53.21
|
0.85
|
1.40
|
T04
|
0.4633
|
0.17726
|
41.9
|
0.90
|
1.26
|
Correlation coefficient with Average Permeability
|
-0.999
|
0.1723
|
0.6884
|
-0.3963
|
Correlation coefficient with CO mass fraction
|
-
|
-0.2993
|
-
|
-0.8344
|
Case 2
|
Average Permeability
|
Frontal Area Density
|
Average Building Separation
|
Average readings at street level of 1.5m
|
Tower with block podium
|
|
|
|
CO mass fraction
|
Velocity (m/s)
|
TP01
|
0.4451
|
0.1014
|
32.47
|
0.74
|
1.76
|
TP02
|
0.2694
|
0.1331
|
25.6
|
0.75
|
1.84
|
TP03
|
0.5971
|
0.07307
|
20.8
|
0.83
|
2.17
|
TP04
|
0.4105
|
0.1079
|
15.13
|
0.81
|
2.01
|
Correlation coefficient with Average Permeability
|
-0.999
|
-0.1649
|
0.6705
|
0.6834
|
Correlation coefficient with CO mass fraction
|
-
|
-0.8275
|
-
|
0.9793
|
Case 3
|
Average Permeability
|
Frontal Area Density
|
Average Building Separation
|
Average readings at street level of 1.5m
|
Tower with Tiered podium
|
|
|
|
CO mass fraction
|
Velocity (m/s)
|
TTP01
|
0.4562
|
0.1124
|
32.47
|
0.82
|
1.94
|
TTP02
|
0.2803
|
0.1687
|
25.6
|
0.80
|
2.19
|
TTP03
|
0.6195
|
0.0896
|
20.8
|
0.92
|
1.77
|
Correlation coefficient with Average Permeability
|
-0.9762
|
-0.3895
|
0.9253
|
-0.9961
|
Correlation coefficient with CO mass fraction
|
-
|
-0.7096
|
-
|
-0.8884
|
Due to the block podiums in Case 02 a channeling effect was seen within the canyon, which resulted in higher wind speeds even though the permeability of building facades was low. However, especially within the podium height, the stagnation of pollutants was high (figure 4). This resulted in a positive correlation between CO mass fraction and velocity.
On the contrary, in Case 03 a strong negative correlation was seen between permeability and wind velocity. As shown by the streamlines (figure 4) the tiered podium pushed pollution along the face of the podium which enables higher dispersion than a block podium. As wind speeds increased, the potential for pollution dispersion from the street level was thus increased.
The assessment zone to calculate permeability was based on the effective length/face of the adjoining blocks which were not broken by a street. CO mass fraction and velocity readings were correlated with the permeability of corresponding assessment zones of the buildings windward of each probe (Figure 09). First we considered the maximum height of a tower as the assessment zone boundary (Table 08) and thereafter considered a maximum height of 30m as the assessment zone boundary (Table 9). This was done to determine if the permeability of podiums had a higher impact on pollution dispersion at street level than the complete building (podium + tower).
When considering both assessment zones and at all probe locations, scenario 03 showed the highest permeability (T03, TP03 & TTP03), followed by Scenario 04, 01 and 02 showing the least.
Readings from probe G at the buffer zone showed a positive correlation between CO mass fraction and permeability. This suggested that the lesser the permeability factor of buildings, the lower the pollution seeping out into the existing context. At probe G; Case 01 had a higher degree of correlation between CO mass fraction and permeability, followed by 03 and 02. As seen above, block podiums significantly alter wind patterns compared to other built forms which resulting in a lower degree of correlation.
When comparing between the two assessment zones (tower height and height up to the 30m), excluding probe G, all other probes located within canyons showed a varying correlation between CO mass fraction and velocity with permeability. This suggests that other parameters such as canyon geometry, podium form, wind channels and pressure differences due to location and orientation may have a higher impact on pollutant dispersion and velocities at street level than permeability alone within canyons. A further study taking into consideration junctions, wind paths, and buoyancy can be done to determine the cause for such large variations.
As discussed in section 2, ventilation indices such as exchange velocity which are based on data at the urban canopy layer have been used in the past to quantify urban ventilation. From the above study, the effectiveness of using such indices can be questioned. In scenario 02, the pollutant plume dispersed upwards along the canyon, while in scenario 03 it dispersed laterally showing very low concentrations at the urban canopy layer. Using data at the urban canopy layer alone to compare complex urban clusters such as those studied here without considering streamlines would not give accurate results.