Abstract
By 2050, it is expected the average global temperature will have risen by 0.7°C and, assuming a constant rate of fertility, the global population will be 11 billion, 66% of which will be living within an urban environment (United Nations, 2014). Both of these factors will exacerbate Urban Heat Islands (UHI) around the globe, increasing energy consumption and potentially putting lives at risk.
Several laboratory experiments were conducted to quantify the cooling effects of water upon an urban surface. It was found that if the surrounding air was at least 20°C cooler than the urban surface in question, pouring 7.5L per m2 of water (≈17°C) significantly cooled the urban material. A field experiment was also undertaken to assess the effects of rainfall on the temperatures of a rural and urban stream. An energy transfer between the rainwater and the urban surface was identified which indicated surface water could be a viable means of transporting heat out of a UHI. Rainwater attenuation and grey water harvesting were identified as being a sustainable source of water as they make use of an otherwise wasted resource.
It was concluded that using water to cool the streets of a city could be a feasible way of cooling UHI temperatures, which could dramatically decrease a cities energy consumption and thus, reduce emissions and anthropogenic heat generated.
2.1 What is an Urban Heat Island?
2.2 How does Climate Change affect UHIs?
2.3 How does Climate Change affect Energy Consumption?
2.4 Temperature Differential between Urban and Rural Streams
4 Possible Surface Water Cooling Techniques
1 Introduction
1.1 Context
Between 1960 and 2015, the global urban population rose from 34% to 54% of the world’s total population (The World Bank, 2016) and is expected to rise to 66% by 2050 (United Nations, 2014). Cities are having to expand in size to accommodate this increase in urban population. Coupled with climate change and subsequent global warming, more prominent ‘Urban Heat Islands’ (UHI) are being formed, where higher temperatures are found within an urban environment than in its surrounding natural land cover (Shahmohamadi, et al., 2011).
Climate change has led to an increase in average global temperature (Section 2.2), which has resulted in even higher temperatures being found within UHIs. This puts the urban population at a greater risk of heat related illnesses, such as heat stroke and exhaustion (Shahmohamadi, et al., 2011). It also leads to excessive energy consumption used for space cooling, such as air conditioning and fans, which increases the amount of emissions and anthropogenic heat generated. This then exacerbates the conditions within the UHI further, creating a ‘snow-ball effect’ situation.
1.2 Research Question
Surface water within the urban environment is regarded as a nuisance and as such, cities have implemented strategies to remove it as quickly as possible. But is this a waste of potential resource?
When rainfall occurs on a warm day, energy (in the form of heat) is transferred from the ground to the rainwater. This decreases the Earth’s surface temperature, resulting in a reduction of the air temperature directly above it. The question that is explored in the following thesis is therefore as follows:
“Could these cooling effects be artificially replicated in a UHI to cool the air down? If so, what would be the implications on energy consumption and resulting emissions within a city?”.
Potential sustainable water sources have been briefly explored.
1.3 Aims and Objectives
This study has aimed to assess the feasibility of using water to cool down streets within a city. The relationship between the surface and air temperature has not been assessed but it has been assumed that by cooling the urban surface, the air directly above would have also cooled.
To achieve this aim, the following objectives were defined:
- To perform several laboratory experiments from which the cooling effects of water upon an urban surface can be quantified.
- To conduct a field experiment comparing the temperatures of a rural and urban stream after rainfall.
2 Literature Review
This Literature Review aims to:
- establish the main causes of UHIs;
- assess the relevant effects of Climate Change;
- identify any relevant research that has already been done; and
- investigate methods and theories that may be of use when analysing the experimental results.
2.1 What is an Urban Heat Island?
Typically, a UHI is formed when a large area of natural land cover is replaced with man-made, impermeable areas. An increase in temperature occurs as a result of several different property changes which, have been summarised below.
| Figure 2.1: Graphic showing the variation in air temperature of a UHI. Modified from the Climate Change Impacts and Adaptation: A Canadian Perspective (2004) |
2.1.1 Porosity of Land Cover
Natural land cover is generally porous and therefore traps moisture, whereas man-made surfaces are comparatively non-porous and therefore cannot. Moisture trapped within a material absorbs radiation from the sun and utilises it through the process of evaporation. This process cools the surface as energy is released as latent heat. Therefore, when natural land cover is replaced with man-made surfaces, the amount of evaporative cooling is significantly reduced, and therefore the surface temperature increases.
2.1.2 Vegetation Removal
Vegetation acts as a natural cooling device, providing shade and evapotranspiration cooling effects. Evapotranspiration is the process by which moisture is transferred from the earth to the atmosphere via evaporation and transpiration from plants (Collins English Dictionary – Complete & Unabridged 10th Edition, 2017).
| Figure 2.2: Relationship between impervious area and evapotranspiration. Modified from The Federal Interagency Stream Restoration Working Group (1998) |
It can be seen in Figure 2.2 that the more impervious material present on the surface, the less evapotranspiration occurs. As this is thought to be one of the dominating surface cooling effects, significantly reducing the rate of evapotranspiration will significantly increase the surface temperature.
Most solar radiation incident upon vegetation is absorbed and used to drive the process of photosynthesis, reducing the amount of radiation being absorbed and released as heat. Since urban surfaces have no vegetation and little to no porosity, the effects of evapotranspiration are minimal, therefore almost all solar radiation incident upon an urban material is absorbed and released as heat.
The effects of convective cooling in UHIs are small when compared to its surrounding rural environment as urban structures have a very small surface area to volume ratio, while vegetation has a very large volume to surface area ration. As UHIs do not lose their heat as readily as vegetation they tend to store energy for longer, releasing more heat at night than their surrounding rural areas. This maintains a high temperature within the UHI which explains why they are most prominent during the night.
2.1.3 Sky View Factor
Buildings in the urban environment are usually constructed in narrow arrangements, reducing the sky view factor (SVF). The SVF is a dimensionless value ranging from 0 to 1 that quantifies the fraction of sky visible from the ground up (Hammerle, et al., 2011). A SVF of 1 means the entire sky is visible. Little solar radiation entering the narrow urban canyon can escape, as it is either absorbed or reflected repeatedly until almost all of it has been absorbed by the urban materials. This has been shown graphically in Figure 2.3.
| Figure 2.3: Graphic showing emission and reflection of solar radiation and anthropogenic heat. Modified from Urban Heat Island Basics (2008) |
2.1.4 Anthropogenic Heat
Anthropogenic heat is the heat released as a result of energy consumption, for example, the heat released from a car. As a large quantity of energy is consumed by the urban environment, anthropogenic heat plays a significant role in the creation of a UHIs.
Research conducted by Salamanca, et al. (2014) concluded that the use of air conditioning within a UHI can increase its surface temperature by up to 1°C at night. This increase in surface temperature results in an increase of energy being consumed for space cooling and therefore increases the generation of anthropogenic heat, exacerbating the UHI further, causing a ‘snow-ball effect’ issue.
2.1.5 Surface Albedos
The albedo of a material is a unitless value quantifying how much incident light hitting a surface is reflected without being absorbed (see Equation [1]).
| Albedo=Lr/Li |
where
Lris the light reflected from the surface and,
Liis the light incident on the surface. A material that appears white will have a higher albedo than a material that appears dark (see Figure 2.4).
| Figure 2.4: Graphic describing the albedo value |
Many materials found within a UHI are relatively dark in colour and therefore have a low albedo value. Therefore, a significant proportion of the solar radiation incident on a UHI is absorbed and then released as heat.
Natural materials, such as trees and grass, usually have higher albedo values compared to urban materials (Figure 2.5). UHIs therefore absorb more radiation than their surrounding rural areas and subsequently have higher surface temperatures, resulting in more prominent UHIs being formed.
| Figure 2.5: Various urban environment albedos (Goodman, 1999) |
2.2 How does Climate Change affect UHIs?
Global warming is likely to increase both the frequency and intensity of heat waves. This will have an adverse effect on UHIs, as the energy consumption required for space cooling will increase, and the risk to urban inhabitants’ wellbeing will become more severe (Tan, et al., 2010).
The term ‘temperature anomaly’ is defined as the difference from the long-term average global temperature (NOAA, 2016). Between 1880 and 2015 the global temperature anomaly rose from -0.19°C to +0.87°C (NASA, 2016), an increase of 1.06°C. The global temperature anomaly and global population have been plotted in Figure 2.6.
Figure 2.6: A graph showing how the global population growth and the global temperature anomaly are changing in a similar manner. Data source: global temperature anomaly data obtained from NASA (2016), early population data up until 1950 is taken from the History Database of the Global Environment (HYDE) (2007), population data from 1950-1960 obtained from The World Bank (2016).
It can be seen in Figure 2.6 that the temperature anomalies started to increase at an accelerated rate in the years leading up to 1980. Between 1980 and 2015 the average annual global temperature increased by approximately 0.2°C per decade. Assuming the average global temperature continues to rise at 0.2°C per decade, by 2050, UHI’s across the globe could have increased in temperature by 0.7°C.
Assuming a constant rate of human fertility, it is likely that the global population will be approximately 11 billion by 2050 (see Figure 2.6); 66% of which will be living in urban environments (United Nations, 2014). This amounts to 7.3 billion people living in cities; a population larger than the entire global population in 2015.
The increase in urban population and global temperature will exacerbate UHIs further, leading to a greater demand in energy required for space cooling as well as putting more people at risk of heat related illnesses and deaths (Shahmohamadi, et al., 2011).
2.3 How does Climate Change affect Energy Consumption?
Figure 2.7 shows a breakdown of the final energy use by sector. The OECD (Organisation for Economic Co-Opertion and Development) Europe, the US, China, and India were studied. Specifically looking at the US, the second largest users of energy were buildings, accounting for approximately 30% of the total energy consumption of the country. Residential buildings account for approximately 54% of this consumption (U.S. Department of Energy, 2012) and the estimated breakdown by end use (the final use of energy) can be seen in Figure 2.8.
| Figure 2.7: Final energy use by sector, 2007 (IEA, 2010) | Figure 2.8: Estimated U.S. residential sector electricity consumption by end use (U.S. EIA, 2014) |
According to the World Bank (2017), in 2014, 81% of the US population lived in an urban environment. It is therefore fair to assume that the electricity consumption by end use shown in Figure 2.8 is representative of urban residential buildings. The data suggests that 13% of the total electricity consumed by residential buildings was used for space cooling. Assuming the global temperature rises by 0.2°C per decade (Section 2.2), by 2050, the global temperature will have risen by approximately 0.7°C since 2015. The electricity demand for space cooling would therefore increase by 5-20% (EPA, 2016) which would dramatically increase the total electricity consumption in the residential sector. Subsequently, emissions and waste heat within urban environments would be escalated, worsening the UHI phenomenon.
2.4 Temperature Differential between Urban and Rural Streams
The temperature of 60 different streams located within regions of varying urban density in North Carolina, USA, were monitored as part of an experiment conducted by Somers, et al. (2013). It was identified that streams influenced by urban development rose by as much as 4°C during a small regional storm, while, for the same storm event, the temperature of rural streams increased by a negligible amount (Somers, et al., 2013). This indicates the surface water runoff entering the urban stream was warmer than the runoff entering the rural stream. It can therefore be concluded that the urban surface warmed the runoff water, meaning the urban surface itself must have cooled.
2.5 Newton’s Law of Cooling
Sir Isaac Newton’s Law of Cooling [2] states that the rate of temperature change of an object is proportional to the difference between the temperature of the body and that of the surrounding environment (University, 2011). This law can be used to calculate the temperature of an object as heat is dissipated into the surrounding environment.
| Tt=Ts+T0-Tse-kt |
where
Ttis the temperature of an object at a certain time (K),
tis the time (s),
Tsis the temperature of the surroundings (K),
T0is the starting temperature of the object (K), and
kis the cooling constant (s-1). This value depends upon the objects surface properties and can be derived through experimental processes.
A model based on Newton’s Law of Cooling produces an exponential decrease over time; the temperature change becomes less as the temperature of the object tends to that of the surrounding environment.
2.6 Least Squares Method
The least squares method is a mathematical regression analysis that finds the line of best fit for a set of data. It aims to create a line that minimises the errors generated between the observed value and anticipated value (Investopedia, n.d.). To prevent negative errors being cancelled out by positive errors, the errors are all squared and summed. The line which returns the least squares value is the line of best fit.
2.7 Conclusion
After reviewing the previously mentioned literature, it becomes apparent that the rate at which UHIs are being formed, and indeed, the rate at which they are increasing in intensity, must be slowed. Several UHI mitigation strategies already exist and are being developed and monitored at this present time. These strategies involve increasing the surface reflectance of urban materials and increasing the vegetation cover within UHIs. These mitigation strategies aim to prevent the UHI from getting too hot, however, if there is a heat wave, they cannot be altered to decrease the temperature more. As the use of water cooling could be controlled, if there was a period of exceptionally high heat, then more water could be poured onto the streets to achieve more cooling. The effects of water cooling will be quantified and the feasibility of using water to cool the streets will be assessed in this thesis.
3 Experiment Overview
The feasibility of using surface water to cool UHIs has been investigated using several laboratory experiments and a field experiment.
A laboratory experiment was conducted to form a relationship between the temperature change of an urban surface and the volume of water poured over it. Environmental conditions, such as relative humidity and room temperature, were controlled and kept constant in the laboratory. This was important as the only variable was the starting temperature of the concrete slab.
A field experiment was also conducted with the intention of identifying a link between rainfall events and the temperature changes of rural and urban streams. Identifying an increase in urban stream temperature after rainfall could suggest surface water is a good medium for removing heat from the urban environment.
3.1 Laboratory Experiment
It was important to understand the effects surface water had on the temperature of urban materials as this directly impacted the feasibility of using rainwater to cool urban streets down. To accurately measure and fully understand the temperature change of an urban material, it was concluded that the results would be best achieved by casting thermocouples into a unique concrete slab of known dimensions, material, density, and mass. The methodology is discussed in detail in Section 3.1.1.
3.1.1 Methodology
Streets in urban environments can be constructed from several different materials; most commonly asphalt and concrete. For the purpose of this investigation a concrete slab was cast to replicate the urban surface.
It was decided the best way of monitoring the temperature of the concrete slab temperature was to use thermocouples. A thermocouple is a temperature-measuring device with two wires, made from different metals, welded together at one end, forming a junction. A voltage is induced in the wires when a temperature change is felt at the junction and this voltage can be interpreted using thermocouple reference tables to calculate the temperature or by using computer software (ThermocoupleInfo, n.d.). For this experiment, computer software developed by Pico Technology was used.
Fifteen type T thermocouples were formed into a 5 x 3 grid ensuring they were evenly spaced to mitigate against any localised cooling because of uneven water pouring. This grid was then cast into a C40 concrete slab. The slab was cast using the dimensions shown in Figure 3.1 as these were representative of a concrete paving slab.
Plan
Elevation
| Figure 3.1: Concrete slab dimensions (all dimensions in mm) |
The concrete slab was placed inside an environmental chamber (see Figure 3.2) and each thermocouple was connected to a data logger which relayed the temperature of the concrete to a computer program called PicoLog Recorder. The environmental chamber was set to the required temperature and relative humidity (RH) of 50% which was kept constant for all of the experiments.
| Figure 3.2: Concrete slab within environmental chamber |
Once the concrete had reached the stated temperature it was removed and placed in an environmentally controlled room, again, with a controlled temperature (≈ 20°C) and RH (≈ 50%).
The thermocouples were plugged into the data logger and set to take temperature recordings every second for 150 minutes while the concrete slab cooled in air. This procedure was then carried out at the same temperature, except this time 1.2L of water (≈17°C) was poured over the concrete surface. It was necessary to pour the water evenly over the concrete to negate localised cooling effects. This was done by using a square edged container attached to a metal bar which could rotate around a central point. The water pouring experiment set up can be seen in Figure 3.3.
| Figure 3.3: Water pouring experiment set up with square edged container on the right-hand side |
To understand how the temperature of the concrete was changed solely because of water pouring it was necessary to record the concrete slab cooling in air and with the addition of water.
The urban surface temperature can vary significantly from region to region and from season to season. For this reason, it was important to repeat the laboratory experiment at several different surface temperatures, all of which could be found within a UHI.
After corresponding with Simone Kotthaus, a Post-Doctoral Research Assistant at the University of Reading, it was concluded that the temperature of an urban surface found within a UHI can range anywhere up to 70°C, with some extreme cases being even higher. It was therefore decided that the laboratory experiment would be conducted at 70, 60, 50, 40 and 30°C and would be named experiment 1, 2, 3, 4, and 5 respectively. All air cooling experiments will be denoted with an A and all water cooling experiments will be denoted with a W, totalling 10 experiments; 1A, 2A, 3A, 4A, 5A,1W, 2W, 3W, 4W, and 5W. No experiments were conducted above 70°C because it is unlikely for the urban surface to reach this temperature or below 30°C as surfaces exhibiting these temperatures would not require cooling as the air temperature above would not be extreme.
3.1.2 Results
A thermocouple was installed in the environmentally controlled room to monitor the room temperature. The room temperature varied slightly so an average temperature for each experiment was calculated. This value was then subtracted from the average concrete temperature to find the temperature difference between the slab and its surroundings. This was done for each second and the results were plotted against time on the x-axis. The air cooling results have been summarised in Figure 3.4.
| Figure 3.4: Graph showing the temperature decay of the concrete slab in air cooling experiments |
Each plotted line within Figure 3.4 represents the slabs respective starting temperature. Experiments 1 and 4 appear to decay exponentially from t = 0s whereas the others appear to decay exponentially from approximately t = 250s. All the exponential decays tend towards T = 0°C however, because of to time constraints, the experiment was ceased after only 6000s, thus T=0°C was never reached.
Using the least squares method, as explained in Section 2.6, the exponential line of best fit was calculated for each experiment respectively. The equation was shown to fit this form:
| y=Ce-kx |
where
Cis the y-intercept (°C), and
kis the cooling constant (1/s). The exponential equation parameters and the coefficient of determination, R2, for the air cooling experiments have been summarised in Table 3.1. The coefficient of determination is a unitless value that ranges between 0 and 1. The closer the value is to 1, the closer the data is to the line of best fit.
| Table 3.1: Table summarising the exponential equation parameters for the air cooling experiments |
| Experiment Reference | C (°C) | k x10-4 (1/s) | R2 |
| 1A | 55.682 | 2.126 | 1.000 |
| 1A* | 55.741 | 2.135 | 1.000 |
| 2A | 43.824 | 1.977 | 1.000 |
| 2A* | 43.924 | 1.993 | 1.000 |
| 3A | 35.399 | 1.802 | 1.000 |
| 3A* | 35.024 | 1.778 | 1.000 |
| 4A | 25.746 | 1.822 | 1.000 |
| 4A* | 25.467 | 1.797 | 1.000 |
| 5A | 8.076 | 3.752 | 0.997 |
| 5A* | 8.946 | 3.987 | 0.996 |
| Note(s):
The experiments marked with a ‘*’ indicate the exponential equations that have been calculated using only data from t = 2000s and t = 6000s. |
|||
The exponential equations calculated using data between t = 2000s and t = 6000s have been superimposed on to assess the well-fitting of the exponential line of best fit (see Figure 3.5).
| Figure 3.5: Graph showing the exponential line of best fit superimposed onto the temperature decay of concrete slab in air cooling experiments |
The dashed lines represent the exponential lines of best fit while the solid lines represent the original data shown in Figure 3.4.
The temperature of the slab above the surroundings temperature has been plotted against time for the water cooling experiments.
| Figure 3.6: Graph showing the temperature decay of the concrete slab in water cooling experiments |
At t = 0s the 1.2L of water was poured over the slab. Each plotted line represents a different water cooling experiment and can be identified from the legend on the right-hand side of the graph. Data beyond t = 5850s for experiments 3 and 4 is missing because of time restrictions which meant that these experiments were not able to run for the same duration as the others.
Using the least squares method again, the equation of the best fitting exponential line was calculated for the water cooling experiments. Assuming the equation has the same form as that shown in Equation [3], the relevant parameters and the coefficient of determination have been stipulated below.
| Table 3.2: Table summarising the exponential equation parameters for the water cooling experiments |
| Experiment Reference | C (°C) | k x10-4 (1/s) | R2 |
| 1W | 47.365 | 2.512 | 0.989 |
| 1W* | 42.212 | 2.364 | 0.999 |
| 2W | 35.327 | 2.022 | 0.991 |
| 2W* | 32.994 | 1.943 | 0.999 |
| 3W | 27.542 | 2.384 | 0.988 |
| 3W* | 24.202 | 2.081 | 0.999 |
| 4W | 21.025 | 3.137 | 0.998 |
| 4W* | 19.398 | 2.950 | 0.999 |
| 5W | 18.809 | 3.607 | 0.998 |
| 5W* | 16.418 | 3.454 | 1.000 |
| Note(s):
The experiments marked with a ‘*’ indicate the exponential equations that have been calculated using only data from t = 2000s and t = 6000s. |
|||
The exponential lines calculated using data between t = 2000s and t = 6000s have been superimposed onto Figure 3.6 to assess the well-fitting of the exponential line of best fit (see Figure 3.7).
| Figure 3.7: Graph showing the exponential line of best fit superimposed onto the temperature decay of concrete slab in water cooling experiments |
The graphs below show how the slab temperature decreased because of the pouring of water (solid line) and the predicted temperature decrease had no water been poured (dashed line). The dashed lines were calculated using data between t = 2000s and 6000s for the relevant air cooling experiment (i.e. 1W was graphed with 1A* as per Table 3.1).
Figure 3.8: Graphs showing the expected air cooling and actual water cooling effects.
From Figure 3.8 it was possible to subtract the expected temperature loss (dashed line) from the recorded temperature loss (solid line) to quantify the sole effects of pouring water onto the slab. These results are summarised in Figure 3.9.
Figure 3.9: Graph summarising the effects of water pouring for each experiment
For experiments 1, 2, 3, 4, and 5, the peak heat losses were 9.0, 6.0, 5.2, 5.1 and 0.3°C respectively occurring at t = 1863s, 1344s, 2060s, 3237s, and 365s respectively.
The experiments summarised in Figure 3.9 involved pouring 1.2L of water (≈17°C) over the 0.16m2 slab. Assuming the same ratio of water to surface area, the volume of water required to have the same cooling effects over a 1m2 slab has been calculated below:
| 1.2L0.16m2=x1m2 | ||
| ∴x=1.2L0.16m2×1m2=7.5L |
From calculation [4] it was identified that 7.5 l of water would be needed to have the same cooling effects as those shown in Figure 3.9 over a 1m2 concrete slab.
Figure 3.10: Graph summarising the effects of water pouring for a specific ΔT
Figure 3.9 has been adjusted to give Figure 3.10 which can be used for a range of different situations. Where ΔT is the difference between the initial surface temperature and the surroundings temperature and the peak times have been rounded to the nearest minute.
3.1.3 Discussion
It is a well-known fact that when an object is warmer than its surroundings, it will cool until a temperature equilibrium has been met. Sir Isaac Newton showed, through the analysis of experiments, that the temperature of an object will in fact change exponentially until the temperature of the object and its surroundings are equal. This is called Newton’s Law of Cooling and the model formula is stated in Equation [2].
Figure 3.4 shows how the temperature of a concrete slab decreases because of air cooling. It appears the temperature change was exponential for all experiments, suggesting that they obeyed Newton’s Law of Cooling. To quantitively show this, it was necessary to see how well an exponential line of best fit suited the datasets. Using the least squares method, the exponential line of best fit and the coefficient of determination, R2, was calculated for each experiment. It can be seen in Figure 3.4 that experiments 2A, 3A and 5A exhibited non-exponential characteristics between t = 0s and approximately t = 500s. Therefore, to give a more accurate line of best fit, the exponential line equation and the coefficient of determination were re-calculated using only data between t = 2000s and t = 6000s to ensure that any initial non-exponential effects were negated. Both sets of results have been summarised in Table 3.1.
Theoretically the cooling constant,
k, is object specific. Since the same concrete slab was used for all experiments, the cooling constant should have also remained the same. However, from Table 3.1 it can be seen that the cooling constant varied between experiments. This is thought to be because of varying temperature differences between the object and its surroundings which can sometimes change the cooling constant (Jiji, 2009).
It can be seen in Table 3.1 that all air cooling experiments, except experiment 5A, exhibit an almost perfect exponential decay in temperature. This quantitively validates the air cooling experimental set up as it shows the results fit the well-founded theory of Newton’s Law of Cooling. A qualitative check can also be conducted on Figure 3.5 to, again, show the exponential equations fit the recorded data. Although experiment 5A exhibited a strong coefficient of determination, it was not equal to 1.000 alike the other experiments. It is thought that this may be because the initial temperature of the slab was similar to that of the room. As a result, the small fluctuations in room temperature had a significant impact upon the temperature change of the slab. This made the temperature decay inconsistent which deviated the results from the exponential line of best fit. Strangely, the exponential line of best fit calculated using the data between t = 2000s and t = 6000s for experiment 5A had a smaller coefficient of determination than the exponential line of best fit that was calculated using all the data. This backs up the above theory that the closer the temperature of the slab is to its surroundings, the less it follows an exponential trend and the more it is affected by fluctuations in the surrounding air temperature.
It can be seen in Figure 3.6 that all experiments, except experiment 5W, initially cooled at a faster rate than air cooling, hence why an exponential function could not be fitted perfectly (Table 3.2). This indicates that an additional term has been added to Newton’s Law of Cooling [2]. To determine this new term, more regression analysis and research would need to be conducted. It is thought that if more water was poured onto the slab the same effects would be seen, but the accelerated rate of cooling would be maintained for longer, significantly decreasing the surface temperature. As the concrete slab was warmest in experiment 1W it was expected that the effects of evaporative cooling would stop influencing the slab first and therefore the rate of cooling would decrease first. However, the rate of cooling remained higher than the other experiments throughout the procedure, cooling the slab to 9°C above room temperature before experiment 2W as seen in Figure 3.6. This indicates that pouring water onto a surface greater than 50°C warmer than its surroundings will significantly reduce its temperature. Experiments 2W, 3W, and 4W all seemed to show the same cooling curve, while experiment 5W exhibited a much lower cooling rate curve. A mistake was made in this experiment process as some water was accidently poured onto the slab before recording had been commenced. Unfortunately, because of time constraints, it was not possible to rerun the experiment and therefore it must be considered that the results obtained from experiment 5W may be inconclusive.
It can be seen in Table 3.2 the cooling constant,
k, varied considerably more between experiments than the cooling constants shown in Table 3.1. This is thought to have happened because of fluctuations in water temperature and differences in evaporative cooling. The average cooling constant displayed in the water pouring experiments was greater than the average cooling constant exhibited in the air cooling experiments. This was expected as the water caused additional cooling effects which increased the rate of cooling and therefore cooling constant. However, the cooling constant calculated for experiment 5W was lower than that calculated for experiment 5A. This implies that the slab would cool slower if water was poured onto it than if it was left to cool naturally in air. Since a mistake was made when undertaking this experiment, it is likely that this was the reason for the above result and not an unknown physical phenomenon, however, this should be verified through further experimentation.
It can be seen in Figure 3.7 that the exponential lines of best fit do not fit the data for the first 2000s seconds. This was expected because the exponential line was calculated using data between t = 2000s and t = 6000s when air cooling dominated as the effects of pouring water and evaporative cooling no longer influenced the temperature of the slab. However, experiment 5W followed the exponential line closely, implying water pouring didn’t change the rate of cooling significantly. This is thought to be because little evaporative cooling occurred as the slab temperature was not hot enough, and therefore air cooling dominated.
The solid lines plotted in Figure 3.8 show how the temperature of the slab changed with time in the water pouring experiments. The dashed lines represent the exponential temperature decay expected if air cooling were to continue to act on its own. If no water was poured onto the slab, it is expected that the cooling curve would follow this line. It can be seen that all experiments exhibit an accelerated rate of cooling after water was poured onto the slab, except for experiment 5W. This was because little evaporative cooling occurred as the initial slab temperature was low and similar to that of the room. As expected, experiment 1W showed the greatest cooling effect.
Figure 3.9 summarises the sole effects water had on the slab temperature. All experiments reached a peak temperature loss at different times, after which the temperature losses started to decrease. It is thought that this decrease would be exponential, however, not enough data was collected to prove this.
As expected, experiment 1 had the largest peak temperature loss while experiment 5 had the smallest peak temperature loss. The difference between peak 1 and peak 2 was significantly larger than expected and is thought to have occurred because of a large decrease in evaporative cooling. Peak 5 was considerably smaller than the rest of the peaks and is thought to be because the initial temperature of slab was already similar to the temperature of its surroundings. In experiment 5, the temperature losses because of water are negative between t = 800s and t = 5800s which implies that it warmed the slab back up after it had cooled it. This may have been possible if the water had pooled on the slab, cooling it below room temperature, however, this is unlikely as the water would have warmed up before the concrete slab could have cooled to this extent. It has therefore been assumed that this occurred because of the mistake made when undertaking experiment 5W.
It was expected that peak 1 would occur before the other peaks as the rate of evaporative cooling would’ve been the fastest. However, peak 1 occurred after peak 2 and at the same time as peak 3. This may have happened because the water was poured more slowly in experiment 1W, although this is unlikely to have altered the time to peak as much as it has been. It may have also been possible that more water pooled on the slab in this experiment and this would have led to more cooling for a longer amount of time. It would be useful to repeat this experiment to see if this phenomenon reoccurs in the same way.
In experiment 5, it did not take long for the slab temperature to almost reach equilibrium with the surrounding temperature, thus, the time to peak 5 was very short.
3.1.4 Possible Limitations and Future Adaptations
Of the 15 thermocouples cast into the concrete slab; 13 were fully functional and 2 were not. The 2 thermocouples that did not work correctly gave recordings that varied significantly between each recording, sometimes varying by 200°C+ in a second. For this reason, these 2 thermocouples were removed from the results.
A limitation of the experiments was that the air temperature of the environmentally controlled room fluctuated slightly. This wasn’t significant for experiments 1, 2, 3, and 4, however, for experiment 5 it is likely to have been detrimental. If these experiments were carried out again, more focus would be placed on maintaining the room temperature.
The concrete slab cooled from all sides, whereas an actual urban surface, such as a street, could only cool from one side, its top. It is therefore likely that the results gathered from the laboratory experiments exhibit a faster rate of cooling than would actually be seen in a real-life situation. One way in which this could be mitigated is to repeat these experiments but insulate all sides of the slab except the top.
Only 5 experiment temperatures were carried out for this research. It would be worthwhile conducting more experiments at different temperatures to see how the different relationships vary and how the different temperatures fit Figure 3.10 and to increase reliability.
The water pouring experiments undertaken for this dissertation only used 1.2L of water for a 0.16m2 surface (or 7.5L of water for a 1m2 surface). It would be beneficial to repeat the experiments conducted using a 1m2 slab and pour over 1, 2, 3, 4, 5, and 6L of water to see the effects that the volume of water has on cooling.
After the experiments for this dissertation were conducted, a rough understanding of how a concrete slab reacts to water pouring was achieved. It would be interesting to repeat the experiments already carried out, but on a sample made from asphalt. How the two materials react to water pouring could then be compared for similarities.
The water used for this experiment had a temperature of approximately 17°C as it was assumed this was representative of a likely storage temperature within the urban environment. The temperature of the stored water would change considerably because of seasonal variations as well as geographic variations and therefore it would be beneficial to repeat these experiments using several different temperatures of water. This would give an understanding of how the water temperature effects the temperature change of the material.
In future experiments, it would be worth noting the time at which all surface water has evaporated off the slab. Beyond this time, the only temperature losses would be because of air cooling. It would be interesting to then compare the exponential lines from before and after this point to see any differences.
Figure 3.10 could be used to estimate the cooling effects of surface water upon the urban surface provided ΔT is known. However, this graph does not fully describe the cooling situation as the starting temperature is not considered. For example, a slab may be 50°C and exhibit a ΔT of 15°C. The cooling for this situation could be read from the graph and it would suggest very little cooling. However, this is untrue as ΔT = 15°C was derived from experiment 5, where the slab was only 30°C. Therefore it is likely, because the slab is 50°C, more cooling would occur than the graph suggests as more evaporative cooling would occur. Further experimentation is required to understand how the initial slab temperature dictates the cooling effects.
The concrete slab used was relatively thin and it is therefore likely the cooling effects exhibited by a thicker slab would’ve been less. It would be useful to repeat these experiments using a thicker slab to understand how the volume of urban material effects its rate cooling.
3.2 Field Experiment
The temperatures of two streams, one rural and one urban, within the area of Bath, UK, were monitored for 38 days, from 09/03/2017 until 16/04/2017. The aim was to identify any fluctuations in stream temperature after rainfall, paying particular attention to regions in which the urban stream temperature varied while the rural stream remained constant. The methodology has been discussed in detail in Section 3.2.1.
3.2.1 Methodology
There were several important factors to consider when choosing which streams to monitor and these have been listed below:
- The size and composition of the stream catchments;
- The proximity of the streams;
- The geometry and flow rate of the streams; and
- The accessibility of the streams.
It was necessary for both streams to have a similar size catchment and for one of the streams to exhibit dense urban development within its catchment and for the other stream to have little/no urban development within its catchment. A comparison between the two streams could then be undertaken to see what effects the urban development had on the stream temperature.
It was important for the two streams to be near to one another so they were both subject to the same climate. Having the same air temperature, rate of rainfall and time of rainfall meant that the data from each stream could be compared.
It is likely that the geometry and flow rate of a stream has an impact upon how it varies in temperature. For this reason, it is important that the streams chosen have similar cross sections and flow rates to one another.
Easy access to the streams will be required as they will be visited and monitored on a regular basis.
It was decided that HOBO temperature loggers would be used to monitor the temperature of the streams. A HOBO temperature logger is a small device that can measure the temperature of water to within ±0.2°C (Onset, 2017). These devices can be left in the streams for a long period of time which suited this experiment well.
Two similar size streams were identified and analysed to ensure they were suitable for the field experiment. Both streams were located to the south of Bath, UK; one passed beneath Church Street and the other alongside Lyncome Vale. Ordnance Survey (OS) mapping for the area was downloaded from Digimaps and imported into an AutoCAD drawing. From AutoCAD Civil 3D, it was possible to estimate the stream catchments using the contours data. The topography of the sites and the stream locations and catchments have been shown in Figure 3.11.
Urban stream catchment
Rural stream catchment
Figure 3.11: Graphic showing the watercourse paths, topography, and catchments of the site.
The thick blue lines represent the streams being analysed. The urban stream runs in the north-east direction and has a catchment shown by the red boundary. The rural stream runs northwards and has a catchment shown by the green boundary. The streams meet at their most northerly points within their catchments and merge into the River Avon to the north.
The streams are comparable since they both have similar catchment sizes; the urban stream catchment covers 2.65km2 whilst the rural stream catchment covers an area of 1.80km2. The hatched area represents dense urban development. Approximately 60% of the urban stream catchment is urban development whilst this is significantly lower at only 17% for the rural streams catchment. It was therefore concluded that these two streams fit the criteria stipulated above and were chosen for the field experiment.
It was decided that the best place to monitor the streams would be as far downstream as possible before the two watercourses met. This maximised the catchment area and therefore the surface water run off volume. The HOBO devices were attached to bricks before being placed into each stream to ensure they were not washed away. The rural stream was monitored just off Church Street and the urban stream was monitored just off Lyncome Vale. The HOBO device monitoring locations have been indicated in Figure 3.12 and images of the streams at these locations can be seen in Figure 3.13 and Figure 3.14.
Rural Stream
Urban Stream
Figure 3.12: HOBO monitoring locations
| Figure 3.13: Location of the urban stream HOBO device. | Figure 3.14: Location of the rural stream HOBO device. |
Another HOBO device was used to monitor the air temperature. It was important this device was placed in a location that did not experience direct sunlight and could not be rained on as this would yield inconsistent results.
The HOBO devices were programmed to take temperature readings every 5 minutes and were installed in their respective locations on the 09/03/2017. Because of the lack of rainfall in March, the devices were left in the streams for a prolonged period in an attempt to record a heavy rainfall event in April. The devices were removed on 16/04/2017 after 38 days of recording.
The sites were visited three times throughout this period. On each occasion, the data from the HOBO loggers was extracted and the flow rate of the streams were recorded. To ensure an accurate flow rate recording, three measurements were taken and averaged.
3.2.2 Results
The flow rates have been summarised in the table below.
| Table 3.3: Table summarising the stream flow rates (m3/s) |
| Rural Stream | Urban Stream | |
| 26/03/2017 | 0.0599 | 0.0079 |
| 03/04/2017 | 0.0308 | 0.0081 |
| 16/04/2017 | 0.0576 | 0.0026 |
As shown in Table 3.3, the flow rate of the rural stream was much higher than that of the urban stream.
The stream and air temperatures have been plotted on the primary axis and the precipitation rate on the secondary axis both against time. The rainfall data was obtained from Weather Underground (2017).
Figure 3.15: Graph showing air and stream temperatures and precipitation rate (09/03/17 08:00 – 20/03/17 00:00).
Data recording started at 08h00 on the 9th March. Between the 9th and 15th March the HOBO loggers all recorded the same pattern; warming began at approximately 08h00 and a peak high temperature was reached at around 18h00. The temperature then started decreasing, reaching a peak low temperature at 2200h. The stream temperatures fluctuated by approximately 3°C daily, ranging from 8°C to 13°C. The air temperature fluctuated by similar amounts with two large peaks at 5°C and 17°C occurring on the 13th and 15th of March respectively. Rainfall occurred on the 12th and 13th of March however these events were minor and were not seen to cause any changes in the stream temperature.
Between the 16th and 19th of March the air temperature and stream temperatures remained relatively constant throughout the day and night, at approximately 11°C. A peak low air temperature of 5°C occurred on the 17th of March. Minor rainfall events occurred within this timeframe, however the temperature of the streams seemed largely unaffected.
Figure 3.16: Graph showing air and stream temperatures and precipitation rate (20/03/17 00:00 – 24/03/17 00:00).
Between the 20th and 23rd of March the air temperature was consistently lower than that of the streams, exhibiting a peak low of 3°C. Rainfall on the 20th and 22nd can be seen to affect the temperature of the streams and so they have been analysed in more detail below.
| Figure 3.17: Graph showing the stream and air temperatures and the precipitation rate (20/03/17 00:00 – 21/03/17 06:00). |
It can be seen that the temperature of the rural stream peaked earlier than the urban stream and air temperature. It is thought that this may be because of a time error associated with the rural HOBO device. For the ease of comparison, the rural results were adjusted so that its temperature peaked at the same time as the other recordings. This was done by adjusting the timing of the rural stream forward by 2.5 hours. The rural stream temperature was then adjusted again to make its temperature at time t0 equal the temperature of the urban stream at t0 as this made a comparison between the two peaks easier. This yielded the graph below.
| Figure 3.18: Graph showing how the stream and air temperatures varied after rainfall (20/03/17 00:00 – 21/03/17 06:00). |
It can be seen in Figure 3.18 that the initial 0.8mm/h rainfall event at 03h00 had no effect on the temperature of the stream. However, the 2.0mm/h rainfall event at 13h00 clearly influenced the air and stream temperatures. Initially, the air temperature and stream temperatures peaked at around 12h00 and then started cooling. Approximately 2 hours after the rainfall event the air temperature peaked and then decreased from 11°C down to 10°C in the space of an hour. Approximately 2.5 hours after rainfall the temperatures of both streams can be seen to increase by approximately 0.5°C. The temperature of the urban stream surpasses that of the rural stream for a short period of time. After this the temperature of the urban stream decreased from 11°C to 10°C in the space of an hour. From approximately 18h00 onwards the temperatures of both streams appeared to decrease at the same rate, the urban stream remained 0.5°C cooler than the rural stream.
On the 22nd of March a large rainfall event occurred which had a significant effect on the urban stream temperature. Similar to before, the time at which the temperature of the rural stream peaked was adjusted, in this instance by 3 hours, to produce Figure 3.19.
| Figure 3.19: Graph showing the adjusted stream and air temperatures and the precipitation rate (22/03/17 00:00 – 23/03/17 00:00). |
At 04h00 a 4mm/hr rainfall event occurred, causing the temperature of the urban stream to decrease significantly, from approximately 9°C to 7.5°C in just 1.5 hours. After this peak minimum temperature, the urban stream warmed back up and behaved similarly to the rural stream.
Between the 25th and 29th of March the rural stream temperature peaked approximately 5 hours before the air and urban stream temperatures. This was because the rural HOBO device was faulty and as a result there were issues in recording the time correctly.
Figure 3.20: Graph showing air and stream temperatures and precipitation rate (31/03/17 00:00 – 16/04/17 00:00).
Between the 31st of March and 16th of April little rainfall occurred. It can be seen that the air temperature fluctuated significantly, from 3.5°C to 25.5°C, which considerably influenced the stream temperatures.
3.2.3 Discussion
From Table 3.3 it can be seen the rural stream had a much larger flow rate than the urban stream. This may be because the rural stream had a larger natural water source. It could be possible that the flow rate of the urban stream was lower because most of the surface water and watercourses within its catchment were rerouted and piped beneath the roads to make way for urban developments. As the flow rate of the rural stream was much larger, changes in temperature would have required more energy, therefore it is likely the effects of rainfall upon the rural stream were less significant than those seen in the urban stream.
As mentioned in the results, there was an error with one of the HOBO temperature loggers. An attempt was made to correct the results recorded with this logger, however this limited the accuracy.
Daily fluctuations in temperature occur throughout this experiment. This is because of solar radiation warming up the earth’s surface in the day and the earth’s surface cooling naturally in the night. Some deviations from this pattern can be seen and it is thought that this is either down to unique weather conditions or rainfall.
On the most part, the air temperature fluctuated far more than the stream temperatures. As would be expected since water has a much larger specific heat capacity than air, therefore requiring significantly more energy to heat.
Throughout the duration of the field experiment it appeared as if the temperature of the rural stream was almost always higher than that of the urban stream. There are many reasons explaining why this may be the case. There could have been an underground water source feeding cold water into the urban stream just upstream of the urban monitoring location. The urban stream could have been more shaded than the rural stream meaning it would have received less solar gains. The rural stream may have spent more time above ground, making it warmer than the urban stream as it would’ve been more influenced by solar gains.
Since the field experiment was conducted in the cold months of March and April, the urban surfaces themselves were cold. Initially the aim was to identify any increases in temperature after rainfall, however, since it was more likely that the rainwater would cool down when running off an urban surface, it was decided that sudden decreases in temperature should also be identified and analysed.
As shown in Figure 3.18 the temperature of both streams increased approximately 2.5 hours after the 2mm/h rainfall event but then sharply decreased after a further 2 hours later. This could suggest one of two things; either the rainwater was warmed by the surface and upon entering the stream it caused an increase in water temperature; or, more realistically as the effects were far more significant, the rainwater was cooled down by the surface and when it finally made its way into the stream it cooled the stream down, causing the sudden decrease in temperature shown. Both suggested causes indicate that surface water influences the surface temperature, as energy is either gained, causing the rainwater to warm up, or lost, causing the rainwater to cool down. It is observed that the urban stream temperature varied more significantly than the rural stream temperature which indicates that urban surfaces have a larger impact on water temperature than rural surfaces. It could also mean the urban runoff is much quicker and therefore temperature effects are concentrated and therefore larger.
It can be seen in Figure 3.19 the 4mm/h rainfall event was quickly followed by a significant drop in urban stream temperature. Since this event occurred in the early hours of the morning and the air temperature was low, it is likely that the urban surfaces were cold. Since the rural stream was subject to the same rainfall event but did not exhibit a significant drop in temperature, it can be concluded that the urban surface must have cooled the rainwater before it was discharged into the stream. This confirms the urban surface temperature has a direct impact upon the runoff temperature and this would be true vice versa (if the urban surface was hot it would warm the water up).
3.2.4 Possible Limitations and Future Adaptations
When analysing the data, it was noticed that the time shown on the rural HOBO device lagged. It was established that every 5 minutes the logger would fall approximately 10 seconds further behind the actual time. This led to significant differences in times between the devices. An attempt was made to correct this time issue; however, it cannot be fully guaranteed it was implemented correctly.
This experiment relies largely upon heavy rainfall. Unfortunately, there was only one moderately heavy rainfall event so it was difficult to establish whether the effects seen were entirely caused by rainfall rather than some other external stimuli.
The field experiment also depends heavily upon the urban surfaces being hot. Since the experiments were conducted in March and April, the urban surfaces were relatively cold. This meant that runoff did not gain any heat from the urban surface and so little increase in stream temperature was seen.
It is likely that most of the rainwater landing on the urban surface was drained through surface water sewer infrastructure and never actually reached the urban stream. Therefore, the surface water could have warmed up, but it never reached the relevant watercourse and therefore, went undetected.
In the discussion, it has been assumed that a temperature drop after rainfall was because the urban surface cooled the rain. However, the initial temperature of the rainfall could already be cold without being cooled by the urban surface and therefore the stream temperature would drop but not necessarily as a result of the presence of urban development.
The difference in flow rates between the two streams made it difficult to compare them as their magnitudes and responses were very different even though they were subject to the same rainfall event.
The urban catchment was approximately 0.85km2 larger than the rural catchment. This meant more rainfall was collected in the urban catchment, potentially leading to more noticeable effects being observed in the urban stream.
To conclude, if this experiment was to be conducted again it would be necessary to monitor the streams in the warmer and wetter months of the year using a higher specification of HOBO temperature loggers. It is also recommended that streams exhibiting similar flow rates are chosen.
4 Possible Surface Water Cooling Techniques
An example of how the data shown in Figure 3.10 could be applied to a real-life situation has been shown in Section 4.1. There are a number of limitations to consider when using this graph and these have been discussed in Section 3.1.4. The purpose of the system is to cool the air down by cooling the surface below it down. Since the air temperature above the surface will not immediately decrease, it is important to maintain a reduced surface temperature.
4.1 Urban Cooling Example
A concrete pavement located in Dubai is 3m wide and 10m long. The surface temperature, ST, and surrounding air temperature, sT, are approximately 80°C and 40°C respectively. Assuming it is a dry night with little to no wind, the following cooling system has been proposed to decrease the surface temperature by approximately 10°C and by doing so decrease the air temperature above it.
The initial temperature difference between the urban surface and the surrounding air temperature would be:
∆T0=ST0-sT0=80℃-40℃=40℃
The urban surface area is:
Surface Area=3×10=30m2
Assuming the water can be stored at approximately 17°C, the cooling shown above can be achieved if 7.5L of water is poured for every 1m2 of urban surface. This equates to a total of 225L of water for each pour.
7.5L ×30=225L
The water should be poured evenly over the 30m2 pavement so that the surface is cooled evenly. Assuming the surface behaves as shown in the graph above, after 20 minutes, the surface temperature will have decreased by approximately 6°C, making the new surface temperature,
ST1, 74°C. Assuming the surrounding temperature has not changed significantly, the new
∆Twould be:
∆T1=ST1-sT0=74℃-40℃=34℃
If another 225L of water was then poured onto the surface, the temperature would drop by a further 5.2°C after 34 minutes as shown in the graph above. This would make the new surface temperature 69°C; 11°C cooler than
ST0, satisfying the stipulated decrease of 10°C.
As the air temperature above the surface will not immediately decrease, it is important to maintain a reduced surface temperature. For this reason, it is proposed 225L of water is evenly poured over the concrete surface every hour for 2 hours, roughly maintaining it at 70°C. After this amount of time it is likely the air temperature will have had sufficient time to cool and will likely remain this temperature or get even lower until sunrise the next day. It was not possible to ascertain the extent to which the air will cool, but this could be identified with/in future research.
The total volume of water required has been calculated:
initial hour+225L+225L=225L+225L+225L+225L=1000L
It can therefore be seen that 1m3 of water storage would be required for this system to work on a 30m2 surface. This can be scaled up, for example, if the concrete surface was 120m2 then 4m3 of water storage would be required, and so on.
5 Conclusion
As stated previously, the global urban population is increasing, meaning urban environments are become larger and denser. This, coupled with the effects of global warming, has, and will lead to UHIs becoming increasingly intense. This will therefore increase the energy required for space cooling which would further exacerbate the UHI. This thesis has assessed the feasibility of using water to cool the UHI down through the process of laboratory and field experiments.
It was shown in Section 3.1.2 that the laboratory experiments all exhibit, partial, or full, exponential cooling. This implied they obeyed Newton’s Law of Cooling and therefore validated the experiment setups and legitimised the results gathered. The effects of pouring 7.5 l of water over a 1m2 surface were quantified and summarised in Figure 3.10. From this graph, it is possible to estimate the likely cooling effects water will have upon the urban surface provided the difference in temperature between the urban surface and its surroundings, ΔT, is known.
It was identified that if the urban surface was less than or equal to 15°C warmer than its surroundings, the effects of water cooling were negligible. However, if the urban surface was greater than or equal to 20°C warmer than its surroundings, the water cooling effects were significant. It is most likely this cooling system would be operated at night, when the effects of solar radiation cannot warm the surface back up again. At this time, the temperature of the surface would be significantly higher than that of its surroundings and therefore water pouring would have significant cooling effects.
The laboratory results proved that pouring cool water over a warm urban surface can significantly reduce its surface temperature through the transfer of heat energy to the water. An attempt was made to show this in the field by demonstrating that rainwater landing upon an urban development will warm up. Unfortunately, because the field experiment was carried out in the cold months of March and April, the urban surfaces were cold. From the field experiment results it was observed that the urban stream cooled down after one particular rainfall event while the rural stream did not. This suggests the urban surface cooled the rainwater, proving a transfer of energy took place. This could imply that if the urban surface was hot, the water would warm up instead, taking energy away from the urban surface, similar to the experiments conducted in the laboratory. This indicates a link between the field and laboratory experiments, however, further field experimentation is required to prove this link.
If a system, such as the one discussed in Section 4.1, was implemented on a larger scale, it could reduce the energy consumption required for space cooling within a city, thus reducing emissions whilst saving money and electricity. The UHI could potentially even cool as a result of less anthropogenic heat being produced and this would further reduce the need for space cooling continuing a cycle of heat reduction.
However, it is possible that such a system may induce a risk of flooding if the storage tank were to be damaged, but measures could be put in place to mitigate against this, such as a backup drainage system. Another issue could be that the stored water becomes stagnant/ contaminated and odorous if kept for a long duration of time within a warm environment. For this reason, it would be suggested that the stored water is discharged over the street or into the nearest surface water sewer at least once every 4 weeks. Another major disadvantage of this system is that it relies on rainfall. If there is no rain for a long duration of time then no water would be stored to cool the streets. It would therefore be recommended the feasibility of greywater harvesting is assessed before implementing any strategy. The start-up costs for such a system would be significant as the required infrastructure would need to be installed and any important services/infrastructure would need to be re-routed. It is also likely the water would heat up in the storage tanks and so it is recommended the water is kept within insulative storage tanks underground to reduce the effects of warming. The effects of cooling identified in Section 4.1 do not consider the solar heat gains felt by urban materials. It is therefore likely that the cooling effects in the day are far less than at night, for this reason it is suggested such a system be used predominantly at night.
If this research was to be continued it is recommended that the future adaptations shown in Section 3.1.4 and Section 3.2.4 be considered for the laboratory and field experiments respectively. It would be worthwhile completing the field experiment within a warmer, wetter, and more densely urbanised location. An assessment on the feasibility of using grey water harvesting to supply the water cooling system would also be recommended.
After analysing the laboratory and field experiment results, it can be confirmed that using stored water to cool the streets of a city down is a feasible way of reducing the UHI intensity. Both surface water attenuation and grey water harvesting could be implemented to provide a sustainable and cost-free source of water to the storage tanks and although there are a number of potential disadvantages associated with this system of cooling, it is necessary now, more than ever, new sustainable methods of reducing energy consumption are conceived to reduce the effects of and on climate change in the near and far future.
6 Reference List
Ahmed Memon, R., Leung, D. & Chunho, L., 2007. A review on the generation, determination and mitigation of Urban Heat Islands. Journal of Environmental Sciences, Volume 1, p. 122.
Brown, R. G., 2004. First Law of Thermodynamics Summary. [Online]
Available at: https://www.phy.duke.edu/~rgb/Class/phy51/phy51/node59.html
[Accessed 15 April 2017].
Collins English Dictionary – Complete & Unabridged 10th Edition, 2017. Evapotranspiration. [Online]
Available at: http://www.dictionary.com/browse/evapotranspiration
[Accessed 05 March 2017].
EPA, 2008. Urban Heat Island Basics. Reducing Urban Heat Islands: Compendium of Strategies, 1(1), pp. 1-22.
EPA, 2016. Climate Change and Heat Islands. [Online]
Available at: https://www.epa.gov/heat-islands/climate-change-and-heat-islands#ref5
[Accessed 18 December 2016].
EPA, n.d.. Measuring Heat Islands. [Online]
Available at: https://www.epa.gov/heat-islands/measuring-heat-islands
[Accessed 15 April 2017].
ETSAP, 2012. Space Heating and Cooling, s.l.: Energy Technology Systems Analysis Programme.
Georgia State University, n.d.. Global Surface Temperature. [Online]
Available at: http://sites.gsu.edu/geog1112/global-surface-temperature/
[Accessed 23 March 2017].
Goodman, S. J., 1999. Heat Island. [Online]
Available at: https://weather.msfc.nasa.gov/urban/urban_heat_island.html
[Accessed 14 April 2017].
Hammerle, M., Gál, T., Unger, J. & Matzarakis, A., 2011. Comparison of models calculating the sky view factor used for urban climate investigations. Theoretical and Applied Climatology, 105(3), pp. 521-527.
History Database of the Global Environment, 2007. Population. [Online]
Available at: http://themasites.pbl.nl/tridion/en/themasites/hyde/basicdrivingfactors/population/index-2.html
[Accessed 15 March 2017].
IEA, 2010. Energy Technology Perspectives, s.l.: International Energy Agency.
Investopedia, n.d.. Least Squares Method. [Online]
Available at: http://www.investopedia.com/terms/l/least-squares-method.asp
[Accessed 19 April 2017].
Jiji, L. M., 2009. Heat Convection. 2nd ed. New York: Scientific Publishing Services Pvt. Ltd..
Magee, N., Curtis, J. & Wendler, G., 1999. The Urban Heat Island Effect at Fairbanks, Alaska, Fairbanks: Springer-Verlag.
Memon, R. A., Leung, D. Y. & Liu, C.-H., 2009. An investigation of urban heat island intensity (UHII) as an indicator of urban heating. Atmospheric Research, 94(3), pp. 491-500.
NASA, 2016. Global Climate Change: Vital Signs of the Planet – Global Temperature. [Online]
Available at: http://climate.nasa.gov/vital-signs/global-temperature/
[Accessed 18 December 2016].
NOAA, 2016. Global Surface Temperature Anomalies. [Online]
Available at: https://www.ncdc.noaa.gov/monitoring-references/faq/anomalies.php
[Accessed 18 December 2016].
Onset, 2017. HOBO TidbiT v2 Water Temperature Data Logger. [Online]
Available at: http://www.onsetcomp.com/products/data-loggers/utbi-001
[Accessed 10 April 2017].
Salamanca, F. et al., 2014. Anthropogenic heating of the urban environment due to air conditioning. Journal of Geophysical Research: Atmospheres, I(1), p. 119.
Shahmohamadi, P. et al., 2011. The Impact of Anthropogenic Heat on Formation of Urban Heat Island and Energy Consumption Balance. Urban Studies Research, Volume 2011, pp. 1-9.
Somers, K. A. et al., 2013. Streams in the urban heat island: spatial and temporal variability in temperature. Freshwater Science, 32(1), pp. 309-326.
Suryanarayana, N., 1995. Engineering Heat Transfer. Michigan: West Publishing Company.
Tan, J. et al., 2010. The urban heat island and its impact on heat waves and human health in Shanghai. International Journal of Biometeorology, 54(1), pp. 75-84.
The Federal Interagency Stream Restoration Working Group, 1998. Stream Corridor Restoration: Principles, Processes, and Practices, s.l.: s.n.
The World Bank, 2016. Urban Population (% of total). [Online]
Available at: http://data.worldbank.org/indicator/SP.URB.TOTL.IN.ZS?end=2015&name_desc=false&start=1960&view=chart
[Accessed 18 December 2016].
ThermocoupleInfo, n.d.. What is a Thermocouple?. [Online]
Available at: http://www.thermocoupleinfo.com/
[Accessed 26 March 2017].
U.S. Department of Energy, 2012. Buildings Energy Data Book: Residential Sector. [Online]
Available at: http://buildingsdatabook.eren.doe.gov/ChapterIntro2.aspx
[Accessed 18 December 2016].
U.S. EIA, 2014. How is electricity used in U.S. homes?. [Online]
Available at: https://www.eia.gov/tools/faqs/faq.cfm?id=96&t=3
[Accessed 18 December 2016].
UCAR, 2011. Urban Heat Islands. [Online]
Available at: https://scied.ucar.edu/longcontent/urban-heat-islands
[Accessed 18 December 2016].
Underground, W., 2017. Lower Weston ISOMERSE15. [Online]
Available at: https://www.wunderground.com/personal-weather-station/dashboard?ID=ISOMERSE15#history/s20170321/e20170321/mdaily
[Accessed 17 April 2017].
United Nations, 2014. World Urbanization Prospects: The 2014 Revision, Highlights, s.l.: United Nations.
University, A., 2011. Newton’s Law of Cooling. [Online]
Available at: http://vlab.amrita.edu/?sub=1&brch=194&sim=354&cnt=1
[Accessed 18 April 2017].
Warren, F. J., 2004. Climate Change Impacts and Adaptation: A Canadian Perspective, s.l.: Government of Canada.
