Results and Discussion

A total of 500 polygons were counted for the generalized categories. On this basis, the minimum sample size for each stratum were calculated (Table 1).

Table 1. List of strata, weights and minimum sample size per stratum

Zoning Count Wg ng+

C                   89      0.178 36
CD                202      0.404      81
RS                  54       0.108      22
RT                  61        0.122        24
RM                52      0.104    21
I    42    0.084        17

+ rounded

Table 2 summarizes the results of t test for random sample of residential plots. Significant difference (p <0.01) in average income observed between the plots of East Side and West Side. As seen in Figure.2 the population income varies considerably in West Side compared to East Side. This characteristic is captured in the random samples of the residential areas. East Side has a much lower sample variance compared to the West Side.

Table 2. Summary of two sample t test assuming unequal variances for

the difference of average income (F1584)

East Side West Side

Mean                                     26900.67               48341.27
Variance                          41396175.33           287117656.8
Observations    30    30

df 37
t Stat -6.479
p(two-tail) < 0.01

List of plots with their geographic coordinates and DAUIDs are available upon request

1/0

Results

Discussion

A framework to estimate PLAD for the entire city

After taking into account constraints such as hard surfaces, quantification of the field measurements across the sampling frame can lead to an estimate of the potential leaf area density for the East Side Vancouver. To obtain an estimate of canopy cover or PLAD across the entire city, the current plan can be upgraded to a multistage sampling plan to include West Side.

Sampling alternative

Here random plots were generated in reference to polygon centres. If a completely randomized design is used, it is possible to achieve a higher dispersion of the sampling plots, maximizing the coverage of the sampling units throughout the study area. However, with the pre-existence of zoning plans, it seems more appropriate to use a stratified sampling design. An advantage of this approach is that an estimate will be obtained for each stratum and strata can be studied individually and compared between different sides of the city along with factors influencing their planning.

Field work obstacles

To provide an estimate for PLAD, ideally all the sampling units in the frame should be visited. However, plots can be inaccessible for various reasons. For instance, permission to visit private properties might not be obtainable. Replacement plots should be considered as an option in implementation of the field sampling.

Inclusion of additional variables

The primary objective of this project was demonstrating a sampling design that can be applied at the ground level to estimate canopy cover and potential leaf area density. However, it was desirable to investigate whether the established plots can be used to study additional variables in relation to canopy cover and plantable space.

Using socioeconomic data it was shown that the differences in family income can be tested in relation to vegetation cover or plantable area. The results can be taken to indicate that residents desire for greener surroundings can be influenced by their level of income.

Nevertheless, the most meaningful variables to be studied in relation to urban canopy cover (or the potential to increase it) are perhaps environmental variables, particularly air contaminant emissions. For this purpose, resources such as DataBC, and MetorVancouver websites were queried. However, GIS environmental data at the scale of this study was not obtained within the time frame of writing this report.

Conclusion

A stratified random sampling plan was proposed and minimum sample size per stratum was determined for ground level estimation of potential leaf area density for East Side Vancouver. The usability of the plots to incorporate additional variables was also demonstrated. The next stage is visiting the plots using the plot centre coordinates and measuring the existing tree cover and plantable space at each plot. This sampling plan can be extended to a multistage plan in which East Side and West Side can be considered as primary classes with various strata within each.

Fig. 4 Quantification of plot areas

Cells within each plot can be filled corresponding to tree cover or plantable space.

1/1

Fig. 1 Sampling units and zoning

Shows map of the distribution of sampling plots on the East Side and those of the West Side over the zoning districts

Fig. 2 Average income

Shows distribution of sampling plots over the map of average income per dissemination areas. Mean: 36000 Standard deviation: 25000

Fig. 3 Plot level

Shows plot level map of an example residential ploygon. The core plots is shown with dark blue centre, and subplots with green centres

1/1