There is no difference in the distribution of marram grass and ling heather on the windward slope of S1 and S2. Only windward slopes are taken into account as they share similar physical conditions (e.g. wind speed).
Collection of Data
Field study was carried out on 30th May 2004 and Studland Heath.
Sampling method used was interrupted belt transects. A measuring tape was laid out perpendicular to the shore. A 0.5X0.5m quadrat was placed to the right of the measuring tape at 20m intervals. Species laid within the quadrat was identified with their percentage cover recorded.
Wind speed, soil pH, soil temperature and angle of inclination were also recorded at every site.
Results
Raw data:
Table 1: Raw data collected on 30th May 04 by own group. Cells are left empty as data is not available
Processed data:
Table 2: Processed data including data collected from different groups and various sites
Analysis
Processed data (shown in table 2) is used in this section to give a more throughout analysis.
 Graphical analysis
As the variable under investigation (percentage cover of species) is discrete instead of continuous, a Histogram cannot be plotted. Other types of graphs are used in this section.
 Bar chart
A bar chart is plotted to represent the coverage of marram grass and ling heather at different sites. (Refer to table 2 for data)
 Kite diagram and Levelling diagram
The levelling diagram shows the landscape of Studland Heath with the 2 sand dunes defined in the diagram. (S1 is from 060m from the shore and S2 refers to 60120m from the shore)
The kite diagram above illustrates the distribution of marram grass and ling heather at different sites.
 Statistical analysis
All numerical values are corrected to 3 significant figures in this section.
 Average (x)
Average coverage of marram grass (xm)
= Σxm / n
= 170/ 7
= 24.3%
Average coverage of ling heather (xl)
= Σxl/ n
= 916/ 7
= 131%
 Standard deviation
Standard deviation (б) measures how far the set of data deviates from the average (x)
Standard deviation of the coverage of marram grass (бm)
= √ [Σ(xm xm)2 / n]
= √ (1540/ 7)
= 14.8%
Counting from the processed data shown in Table 2:
There are 7 counts lay between 1 standard deviation from the mean, i.e. from (9.539.1)%
Percentage of data lies between 1 standard deviation from the mean
= (7/42) X 100%
= 16.7%
There are 42 counts lay between 2 standard deviation from the mean, i.e. from (053.9)%
Percentage of data lies between 2 standard deviation from the mean
= (42/42) X 100%
= 100%
As only 16.7% of the data lies between 1 standard deviation from the mean and more than 95% of data lies between 2 standard deviations from the mean, this set of data is not normally distributed.
Standard deviation of the coverage of ling heather (бl)
= √ [Σ(xl xl)2 / n]
= √ (135000/ 7)
= 139%
Counting from the processed data shown in Table 2:
There are 42 counts lay between 1 standard deviation from the mean, i.e. from (0270)%
Percentage of data lies between 1 standard deviation from the mean
= (42/42) X 100%
= 100%
Since there are more than 6870% of the data is within 1 standard deviation from the mean, this set of data is also not normally distributed.
 Statistical test
Though the data collected are interval, yet, they are not normally distributed. Thus ttest cannot be applied. Only standard error can be calculated and used to determine if the difference between the two sets of data is significant.
Standard error (S.E.)
= √ [(бm2/ n) + (бl2/ n)]
= √ [ (14.8/ 7) + (139/ 7) ]
= √ 22.0
= 4.69%
Difference between the averages
= xl – xm
= 131 – 24.3
= 106.7%
2 X S.E.
= 4.69 X 2
= 9.36%
As the difference between the averages (106.7%) is greater than 2 X S.E. (9.38%), the results are significant, i.e. there is a significant difference between the distribution of marram grass and ling heather in Studland Heath on 30th May 04.
The null hypothesis is therefore rejected. There is a significant difference in the distribution of marram grass and ling heather on the windward slope of S1 and S2.
Evaluation
Problems in sampling:
 Processed data used in the analysis includes data collected by different groups and different sites. However, groups are not working parallel to each other, thus the data may actually be collected at different sites.
 It is not always possible to collect data in 20m interval due to the landscape. Groups need to make adjustment in various situations, e.g. when approaching the top of the dune.
 It is difficult to identify different species due to lack of experience. Different species may also grow together, overlapping each other, thus making it more difficult to be identified.
 Succession in Studland Heath maybe deflected by human activities, e.g. work has been done to prevent the community develops into woodland. This may affect our assumption that the succession is only affected my physical factors.
 Percentage cover of species is recorded instead of the number of plants. Percentage cover maybe affected by various factors, e.g. size of plants together with number of species. Thus percentage cover may not be a good representation of occurrence of species in different sites.
Overcoming the problems:
 A compass can be used to check the direction of groups and ensure all of them are perpendicular to the shore. This solves problem 1.
 By classifying the species into larger groups, it makes identifying the species easier and helps solving problem 3.
 A more throughout sampling should be done to collect a wider range of data, e.g. collect the data in 10m interval. This may minimise the effect of errors in technique.
 Sampling should be repeated on a day with similar physical conditions to improve its reliability.
References:

Plant succession:

Geology of sand dunes:

EN contribution: