Accessible Green Spaces Analysis - A micro-internship project

article written by Laura Dobbs, TVERC micro-internship student

My micro-internship was focused on assessing the biodiversity of volunteer recorded species records across 69 Accessible Wildlife Sites in Oxford City based on six aspects of the data:

  • Taxon groups
  • Species Richness
  • Correlation between species richness and average net annual income
  • Difference between species richness and greenspace type
  • Jaccard Similarity Index

Taxon Groups

In total 47 taxon groups were identified across all sites. I found the most commonly occurring taxon group was Flowering Plants with 6646 observations made. Moths were the second most commonly occurring (2244 observations), followed by Birds (1546 observations). There were ten taxon groups which had less than five recordings made in total across all sites (Table 1).

I recommend that sites should be surveyed for these ten taxon groups to determine whether few observations have been made due to a genuine lack of biodiversity or volunteers are not trained to recognise and record these taxon types.

Table 1 Taxon Groups with Less Than Five Recordings Across All Sites


Species Richness

I noticed that not all species observation records had an abundance value recorded. Therefore, I was limited to calculating the species richness value for each site which was done using QGIS and Microsoft Excel. Figure 1 shows the distribution and species richness values associated with 69 Accessible Wildlife Sites in Oxford: the sites with the highest species richness values are Sydlings Copse (1704), Shotover Country Park (1335) and Bagley Wood (1249).

Analysis showed eight sites had no recorded species, their locations can be seen in Figure 2. Additionally, nine sites were identified with less than five species records (Figure 3). It is unclear whether these findings are due to a lack of biodiversity, or the sites were not visited frequently enough to capture a more accurate representation of the species present. I recommend that volunteers are encouraged to visit these sites to take recordings for at least species presence so that more information can be gathered about their biodiversity.

Figure 1 Map of species richness across 69 Accessible Wildlife Sites in Oxford city.

Figure 2 Eight sites with no recorded species.

Figure 3 Nine sites with less than five species records each.

Correlation between species richness and average net annual income

I used the species richness data already calculated and found the average net annual income data for different areas of Oxford city. Since the data was not normally distributed, I performed a Spearman’s Rank Correlation in R Studio version 1.0.15. There was not a statistically significant correlation between species richness and average net annual income (rho=0.215, p=0.1).

Difference between species richness and greenspace type

Across Oxford city six different greenspace types were identified, each with a differing number of sites (Table 2).

As the species richness data was not normally distributed, I used a non-parametric Kruskal-Wallis test which showed a statistically significant difference between species richness and the type of greenspace (X2=19.932, p=0.001). A post-hoc Mann-Whitney-Wilcoxon test revealed there was only a significant difference between the species richness values of Nature Reserves and Park and Gardens (p=0.047). As shown in Table 2 there is variation in the number of sites of each type of greenspace. This may have an affect on the validity of the results as there is only one woodland site, whereas there are 24 Parks and Gardens sites.

Table 2 Mean Species Richness and Number of Sites for Each Greenspace Type

Jaccard Similarity Index

A Jaccard Similarity Index calculates the similarity of sites based on the presence and absence of data between sites. A value close to one means the sites are very similar in composition, whereas a value close to zero indicates high dissimilarity. The index was calculated using both taxon group and species data in Microsoft Excel and R Studio.

Taxon Groups

I found the highest similarity score was between the sites ’Arable Land at Marston’ and ‘Pixey Mead Central’ with a score of 0.8 when using taxon groups. Table 3 shows the three highest similarity indexes.

Table 3 Highest Jaccard Similarity Indexes for Taxon Groups


Florence Park and Grandpont Park, Whitehouse Road were the only two sites showing similarity when using the species found (0.54) and are located in different parts of Oxford (Table 4).

Table 4 Highest Jaccard Similarity Index for Species

Based on my findings, sites which have either a low or no species richness score should be surveyed by volunteers to assess the taxon and species types found at the sites. This would determine whether these scores are due to a lack of biodiversity or under sampling. Additionally, it would be beneficial for abundance to be measured by volunteers so that more complex analysis can be undertaken to calculate biodiversity indexes such as Simpson’s and Shannon’s Biodiversity Indexes.

Posted: February 8, 2021