Detailed Methods Add Info

A 0.5 m² quadrat (a square frame measuring 0.5m on each side) was placed on the ground every 2m along the line. The plant species inside the quadrat were identified, their % cover was estimated and their maximum height noted.

This type of survey is called a belt transect, because it measures species abundance across a narrow band (similar to a belt). In addition to the belt transect, a general survey was also carried out to note other plant species occurring outside the transect area.

Progress of the Heathland Restoration Project has been monitored on an annual basis since this initial survey was carried out. The survey in 1996 was based solely on the belt transect method. Subsequent surveys have incorporated random sampling, as well as a belt transect. The belt transect position was not permanently fixed, although it was carried out in virtually the same area each year.

Random quadrats were included from 1997 onwards because it was felt that the belt transect only covered a very narrow band across the centre of the site. While this allowed comparisons to be made between the vegetation of different sections, it did not necessarily accurately reflect the overall plant cover of the sections.

Surveying the Heathland Restoration Project site.

Random quadrats were sampled in each of sections 1 - 9, at the same time as the belt transect was carried out. The position of random quadrats was determined using a random number table to select the number of paces to be walked, on a co-ordinate basis.

As there also appeared to be a visual difference between the vegetation in the top and bottom halves of some of the sections, these were sampled separately to see if there were any measurable differences. The transect line acted as the median point.

For the belt transect method, quadrats were surveyed every other metre along the line in each section. An equivalent number of random quadrats were sampled in the top of each section. This was repeated for the bottom section of each strip. As a result, twice the number of random samples were taken in each strip overall compared to the belt transect.

The two sets of results together cover all areas of the heathland site and give a good picture of alterations in the vegetation cover of the site over time.

Annual surveys were carried out in the same month each year (July) to avoid introducing sampling errors caused by seasonal changes in vegetation. For example, if the project site was surveyed in May when Bluebells and wood-rushes are at their peak in growth and flowering, these species would show as being far more dominant on site than the July surveys indicate. By the time July arrives, both Bluebells and wood-rushes have substantially died back, greatly reducing their percentage cover in sampled quadrats. Other species have then taken over the dominant role for the time being. A second annual survey, this time carried out in July of the following year, would then apparently show that Bluebells and wood-rushes had substantially decreased on the project site over the intervening year. This would clearly be an erroneous conclusion brought about by comparing data which was not comparable.

There will inevitably be seasonal differences in the dominant species in each project strip, therefore it was important to make sure that the surveys were carried out at the same time each year to allow meaningful comparisons to be made. July was chosen because this is the peak time of year for most plant species to be flowering. This greatly aids identification of species, particularly grasses.

Additional Notes

Observer bias:

The same person has conducted all the ecological surveys published here, which means that variations due to observer bias have been avoided.

Use of a 0.5 m² quadrat:

The first survey carried out in 1996 was conducted as a one-time survey. It was not originally intended to carry out the series of annual surveys published here. At the time of the first survey, a 0.5 m²quadrat was entirely suitable for the height range and type of vegetation occurring across the project site. Once this size of quadrat had been selected, it then became necessary to use the same size of quadrat for each subsequent survey in order for results to be directly comparable year on year. (Different sizes of quadrats give different results.)

By 2001, the vegetation in section 5 (the control strip, which has had no management) and also in section 9 (hand weeded) had grown so tall in many places that this quadrat size was no longer at all suitable. However, for the sake of continuity and comparison, the same sized quadrat was still used, with adjustments (the quadrat was treated as a vertical column and everything within the column was noted).

Future surveys will have to address this problem and use a different method of sampling these two sections which will still be as comparable as possible with the others. This illustrates the importance of carefully considering initial survey methods when beginning on a long-running monitoring programme. It is vital to consider how the vegetation is likely to change with time and to employ methods which will be able to encompass the changes envisaged.

Sample of Raw Data Comparison of Sampling Method Results