4.1. Numerical Analysis
The developed ISE models were used to determine the relationship between TDA and the location preferences of urban activities. Usually, households are producers of labor, and these laborers are consumed by services and industries. Likewise, services and industries are the producers of jobs, and these jobs are consumed by households. The location choice behavior of activities was further classified:
- —
Locational preferences of household activities;
- —
Locational preferences of commercial services (household obtained services such as retail).
In principle, most economic activities are dependent on an acceptable level of accessibility in order to survive and develop, so a range of accessibility measures need to be considered. As mentioned earlier, a TDM was developed, validated, and calibrated for the years 2012 and 2015 to calculate TDA to various activity locations. This TDA measure is used to calculate the short-term dynamic access to goods, services, and other activities that are located in the City of Wuhan during the morning peak, off-peak, and evening peak. However, the TDA measure is limited to certain times of the day. The TDA considers only the major modes of transport (car, taxi, bus, and metro), excluding walk and bike modes, which do not influence traffic congestion. Commodity flow from production location to consumption location influences the transport system. The movement of these activities makes the location more or less attractive for households, firms, and businesses. The transport cost coefficients for the commercial services commodities and the labor commodities for the ISE model were based on local data. These coefficients may vary depending on the commodity type that is transported, for instance, it could be different for households and goods. These coefficients are calculated by dividing the vehicle operating cost per unit of the commodity per unit of distance by the money value of the commodity load on the truck.
As mentioned in the methodology, an integrated land use transport system used a composite accessibility approach to calculate the access to goods, services, and other activities that were located in the study area, which needed transport cost coefficients to determine the cost of transporting each commodity type. Furthermore, the composite accessibility measure consists of three major components that are associated with the utility of buying or selling a given commodity in a given zone.
As discussed earlier, the three components of utility play a vital role in driving the decisions in the ISE spatial economic system. These three components influence the location choice behavior of households, firms, and industries. The change in the utility function influences the consumers and producers of the given commodity for households and other activities because these activities are susceptible to commodity utilities. These transportable commodities include goods, services, and labor. The categorization of goods and services is made by grouping the goods and services that are associated with the predefined industries. These all interact with each other through a transport system. Households are social units; usually, they provide labor and consume goods and services, along with space. Meanwhile, industries, firms, and businesses offer and consume goods and services, and they are the consumers of labor and space, as shown in
Figure 7. Usually, households produce labor and these labors are consumed by commercial services. During the process, commercial services consume labor and produce services, and these services are consumed by households in the region.
Table 4 shows the activities and commodities and their production and consumption process. The connectivity between TAZs is based on the congested network of the transport model (time, distance, logsum, etc.), which in transport terms is called disutilities. Disutilities are used to establish the interaction between TAZs for the buying and selling of various commodities. The trips in the system are linked to the economic flow system using the calculated transport coefficients, which reflect the money cost per trip for goods, services, and labor. Due to the lack of truck trip data, the auto is used as a proxy for trucks for transporting goods. Undoubtedly, residents prefer locations with good access to utilities, different services, primary public transport systems, and employment centers.
The primary hypothesis is that residents prefer a location that is close to their place of work, or a place with a high level of access to their primary needs. Resultsrevealed that the actual influence of residential location and accessibility of the desired job location during peak and off-peak periods may depend mainly on the underlying geographical location (spatial) and temporal period (time of the day), as well as the average price of the location.
Overall, the R
2 for the models shows a strong relationship between TDA and residential locations where the adjusted R
2 value was found to be higher than 0.90. The results revealed that there was a significant improvement in the residual square from the years 2012 to 2015, due to the improved level of accessibility in the year 2015, as shown in
Table 5. Furthermore, results showed that household locations were sensitive to morning peak TDA where the R
2 value was found to be higher than 0.90.
Table 6 presents results for the TDA to commercial activities. Overall, the results showed that there was a strong relationship between TDA and commercial locations, where the adjusted R
2 value was found to be greater than 0.91 for the year 2012 and greater than 0.93 for the year 2015. Moreover, the results indicated that TDA to commercial locations had a high R
2 value of 0.98 during off-peak time. Meanwhile, there was a significant improvement in the residual square from the year 2012 to 2015 because of improved TDA during 2015.
Accessibility indexes (logsum) were used to present the TDA levels at each TAZ. For instance, “No” refers to no accessibility; “Low” refers to low accessibility that ranges between 0.22 and 0.92; “low-medium” refers to accessibility levels that are higher than low accessibility and lower than medium accessibility and ranges between 0.93~1.13; “medium” refers to a medium level of accessibility that ranges between 1.14 and 1.28; “medium-high” refers to accessibility that is higher than medium level and lower than high level and ranges between 1.29 and 1.39; and finally, the high accessibility index shows high accessibility in each TAZ and ranges between 1.40 and 1.51 during the years 2012 and 2015.
Meanwhile, the ISE model-estimated location (converted to density) is used to represent the activity locations. For instance, “No” refers to no activity locations; “Low” refers to a low level of activity locations and ranges between 0.003 and 0.004; “low-medium” refers to activity locations that are higher than lower level and lower than medium level and ranges between 0.005 and 0.008; “medium” refers to a medium level of activity locations and ranges between 0.009 and 0.013; “medium-high” refers to an activity location level that is higher than medium level and lower than high level and ranges between 0.014 and 0.021; and finally, the high activity locations index shows high activity locations in each TAZ and ranges between 0.022 and 0.038 during the years 2012 and 2015. These indexes are used to represent the TDA and activity location in each TAZ, as shown in
Table 7. The TDA to household and commercial activities are calculated using the logsum indexes [
75]. The TDA Logsum values are computed using Equation (1), and the activity location density is computed using Equation (3).
4.1.1. Locational Preferences of Household Activities
Households produce labor and are consumed by businesses, firms, and other industries. Except for households, all production, consumption, import, export, and labor wages are defined in terms of RMB (monetary unit). The commuting costs are calculated during the model run, and labor wages are adjusted to match the supply and demand in each location for each occupation. The categorization of employment by industry was made based on occupation in each industry. This study used multi modes (car, taxi, metro, and bus) to calculate TDA for household and commercial activities.
Figure 8a,b indicate the accessibility of household activities during the morning peak for the years 2012 and 2015. The morning peak accessibility of household activities value ranges are low accessibility (0.22~0.92); low-medium accessibility (0.93~1.13); medium accessibility (1.14~1.28); medium-high accessibility (1.29~1.39); and high accessibility (1.40~1.65). The reason for high accessibility in downtown areas is the availability of multi-modes for commuting during morning peak times. In general, when the transport cost increases, the size component (accessible area) for some zones decreases for the consumer of the goods and services. All the exchanges (zones and commodities) become less accessible for the consumers, meaning that the buyers obtain less variety in the product or service. For labor, it is expected that an increase in transport can decrease the average trip length of commuting to work. Benefits losses arise due to the increase in transport disutility, but this also affects the size component, since households lose variety in job options, decreasing their size component (due to the loss in variation).
Figure 9a,b present ISE-estimated household locations using morning peak accessibility for the years 2012 and 2015. The ISE-estimated household location density values using morning peak accessibility ranges are low household locations (0.003~0.004); low-medium household locations (0.005~0.008); medium household locations (0.009 ~0.013); medium-high household locations (0.014~0.021); and high household locations (0.021~0.038). The results also revealed that the downtown area had a high level of TDA during the morning peak. This is mainly attributed to the fact that the frequency of transit during the morning is high to help employees and workers arrive on time to their workplaces.
Figure 10a,b present the accessibility of household activities during the off-peak period for the years 2012 and 2015. The off-peak accessibility of household activities value ranges are low accessibility (0.22~0.92); low-medium accessibility (0.93~1.13); medium accessibility (1.14~1.28); medium-high accessibility (1.29~1.39); and high accessibility (1.40~1.65). Most of the retail services activities are located in the downtown area; during off-peak, the downtown area shows high TDA. The off-peak presents the low level of congestion on transport infrastructure and all activities which are sensitive to time and finds these to be very attractive.
Figure 11a,b present ISE-estimated household locations using off-peak accessibility for the years 2012 and 2015. The ISE-estimated household locations density value ranges using off-peak accessibility are low household locations (0.003~0.004); low-medium household locations (0.005~0.008); medium household locations (0.009~0.013); medium-high household locations (0.014~0.021); high household locations (0.021~0.038); and values that are below 0.001 represent no household locations.
Figure 12a,b present the accessibility of household activities during evening peak for the years 2012 and 2015. The evening-peak accessibility of household activities value ranges are low accessibility (0.22~0.92); low-medium accessibility (0.93~1.13); medium accessibility (1.14~1.28); medium-high accessibility (1.29~1.39); and high accessibility (1.40~1.65). During the evening peak, the level of accessibility of household activities is low compared to the morning peak and off-peak time, as shown in
Figure 12a,b.
Figure 13a,b present ISE-estimated household locations using evening peak accessibility for the years 2012 and 2015. The ISE-estimated household locations density value ranges using evening peak accessibility are low household locations (0.003~0.004); low-medium household locations (0.005~0.008); medium household locations (0.009~0.013); medium-high household locations (0.014~0.021); high household locations (0.021~0.038); and values that are below 0.001 represent no household locations. The household activity locations also dropped significantly because of low TDA for these activities during highly congested periods, as depicted in
Figure 13a,b.
4.1.2. Locational Preferences of Commercial Activities
Commercial services or business service locations are essential in any economic system. Usually, households produce labor and consume goods, services, information, retail hospitality, real estate, business services, technical services, environmental services, and private services. However, an aggregate ‘commercial activity’ term is used to present commercial services activity in the ISE system.
Figure 14 presents the buying and selling process of commercial activities. Commercial activities usually buy labor and sell services to other activities including households. During this process, commercial activities consume various goods and services.
The improvement in the road network and transit services improved the level of accessibility and the location choice behavior of commercial services, with better accessibility attracting more commercial activities. Commercial activities (household obtained services) prefer locations with low transport costs, which means that they only need to pay low transport costs. However, some services are sensitive to morning peak time accessibility, and some services are sensitive to off-peak time accessibility, depending on the type of commercial activity.
Figure 15a,b present the accessibility of commercial service activities during the morning peak for the years 2012 and 2015. The morning peak accessibility of commercial service activities value ranges are low accessibility (0.22~0.92); low-medium accessibility (0.93~1.13); medium accessibility (1.14~1.28); medium-high accessibility (1.29~1.39); and high accessibility (1.40~1.65). The results in
Figure 15a,b present that during the morning peak time, the level of accessibility of commercial activities, especially in the year 2015, is relatively high compared to the 2012 morning peak.
Figure 16a,b present the ISE-estimated commercial locations using morning peak accessibility for the years 2012 and 2015. The ISE-estimated commercial locations density value ranges using morning peak accessibility are low commercial locations (0.003~0.004); low-medium commercial locations (0.005~0.008); medium commercial locations (0.009~0.013); medium-high commercial locations (0.014~0.021); high commercial locations (0.021~0.038); and values that are below 0.001 represent no commercial locations. Meanwhile, the results in
Figure 16a,b also revealed that areas with high TDA to commercial services showed a high density of commercial locations.
Figure 17a,b present the accessibility of commercial service activities during the off-peak period for the years 2012 and 2015. The off-peak accessibility of commercial service activities value ranges are low accessibility (0.22~0.92); low-medium accessibility (0.93~1.13); medium accessibility (1.14~1.28); medium-high accessibility (1.29~1.39); and high accessibility (1.40~1.65).
Figure 17a,b present the TDA to commercial activities during the off-peak period. Results revealed that during 2015 the level of accessibility of commercial activities located in downtown and central business areas of rural zones had a high TDA.
Figure 18a,b present the ISE estimated commercial locations using off-peak accessibility for the years 2012 and 2015. The ISE estimated commercial locations density value ranges using off-peak accessibility: low commercial locations (0.003~0.004), low-medium commercial locations (0.005~0.008), medium commercial locations (0.009~0.013), medium-high commercial locations (0.014~0.021), high commercial locations (0.021~0.038) and values below 0.001 represent no commercial locations. The ISE off-peak time activity location model revealed that commercial activities are susceptible to TDA as shown in
Figure 18a,b. However, some commercial activities prefer and are sensitive to morning peak time accessibility, and most of them are sensitive to off-peak time accessibility. The results revealed that during the off-peak time, the level of accessibility of these services is relatively high compared to the morning peak.
Figure 19a,b present the accessibility of commercial service activities during the evening peak period for the years 2012 and 2015. The evening-peak accessibility of commercial service activities value ranges are low accessibility (0.22~0.92); low-medium accessibility (0.93~1.13); medium accessibility (1.14~1.28); medium-high accessibility (1.29~1.39); and high accessibility (1.40~1.65). The commercial TDA and ISE model results revealed that during the evening peak, the accessibility of most of the household obtained services was low compared to the morning peak and off-peak time, as shown in
Figure 19a,b.
Figure 20a,b present ISE-estimated commercial locations using the evening peak accessibility for the years 2012 and 2015. The ISE-estimated commercial locations density value ranges using evening peak accessibility are low commercial locations (0.003~0.004); low-medium commercial locations (0.005~0.008); medium commercial locations (0.009~0.013); medium-high commercial locations (0.014~0.021); high commercial locations (0.021~0.038); and values that are below 0.001 represent no commercial locations, while values that are below 0.001 represent no household locations. However, during 2015, the level of accessibility of household obtained services increased compared to 2012. This is because, during 2015, new highways were built, new bus routes were added, and a new subway line was put into service which eventually increased the level of accessibility. With increased TDA, in the year 2015, more commercial services moved to high accessibility locations, as shown in
Figure 20a,b. The ISE commercial services model activity locations results revealed that commercial services were susceptible to TDA.