Bert Reid

During 1987 and 1988, the Botanical Society of the British Isles (BSBI) organised a vascular plant survey in selected tetrads throughout the country. Five of these tetrads fell within the area covered by the Worcestershire Flora Project (WFP) - SO95A, SO95J, SO95W, SO98A and SO98W. This exercise was a trial run for the full Atlas 2000 project and the instructions were to attempt to produce a full list of taxa for the selected hectads (SO95 & SO98), with particular emphasis and more detailed recording within the selected tetrads.

BSBI recently set up a partial repeat of this survey during 2003 / 2004, although with significant changes to the survey instructions. The same tetrads were to be surveyed but without any attempt to produce full hectad lists (these had been done for Atlas 2000). A wider range of taxa was to be included in the survey. These included all taxa growing “in the wild”, including planted trees, garden throw-outs and the like. A major aim of the repeat survey was to investigate how far it was possible to use such surveys to assess changes within the flora.

As part of the repeat survey, I agreed to carry out recording within two tetrads during 2003, SO95A and SO95W. To assist this recording I investigated all known previous records for the two tetrads so that I could use these earlier records to help me look for particular taxa. The only localised pre-1970 record I found was in SO95A for Gymnadenium conopsea (Fragrant Orchid) dated 1867 “at Stoulton, in a meadow next the railway”. This record, which was not re-found in 2003, has been ignored in subsequent analysis. Other records were from the 1970s and early 1980s from WWT phase 1 surveys etc., the BSBI surveys from 1987/88, and WFP records from 1989 onwards.

The 1987/88 survey found 255 taxa in SO95A and 249 taxa in SO95W. The corresponding totals in 2003 were 367 and 350. This appears to show a large increase in floral biodiversity (44% and 40%). A more detailed analysis is shown below.

A number of features stand out in this analysis. Many taxa recorded between 1970 and 2002 have not been re-found in 2003 (61 and 50). This could indicate a reduction in biodiversity. Even more taxa were recorded for the first time in 2003 (75 and 68), suggesting biodiversity increase. Less than 2/3rds of all taxa known since 1970 were recorded in the 1987/88 survey (59% and 62%). The 2003 figures are 85% and 87%. A series of factors contribute to this apparent confusion.

The longer that is spent in the field, the more plants will be recorded. The pattern of taxa against time can be shown from the 2003 survey, where time spent in the field was recorded. The relationship is shown graphically below.

The graphs show a clear pattern of steep increases up to about 20 hours followed by a flattening beyond this. Both tetrads appear to be heading towards about 380 - 400 taxa if more time was spent in the field. The instructions for the BSBI 2003 survey were to spend a minimum of ten hours in a tetrad. This is clearly inadequate for any reasonably diverse lowland area. If recording had stopped at ten hours, the totals would have been around 250 in each tetrad - these would have been very similar to the 1987/88 totals.

At the tetrad scale many plants are rare. Even generally widespread plants like Crepis capillaris (Smooth Hawk’s-beard) can be very hard to find. A quick analysis of the 2003 data shows that about 45% of the recorded taxa are known from three or fewer places in a tetrad. Some of these scarce taxa are quite easy to find, especially if they are restricted to scarce habitats like ponds, but others are much more elusive. A handful of plants of Anthemis cotula (Stinking Chamomile) at the edge of one of the many wheat fields is very easily missed. Other plants have only a short recording period, being difficult to identify without flowers or fruits. If they flower very early or late, then visits during the main recording period will miss them.

It is clear that the 1987/88 survey was much less complete than the 2003 one. Less time was spent in the field, and the visits were virtually confined to the main recording period. This makes any change assessment very difficult at the tetrad level. BSBI are well aware of the problems but amalgamating data from across the country may even out some of the inconsistencies. A better comparison for us is between the 2003 survey and all records from 1970 to 2002. These periods have therefore been used for further analysis.

SO95A shows 75 new taxa added and 61 taxa not re-found. The figures for SO95W are 68 added and 50 lost. These gains and losses for each tetrad are discussed separately. The SO95A gains include 13 planted trees and shrubs that would probably not have been recorded in earlier surveys. A further 19 gains are of garden escapes and throw-outs where earlier recording practice would have been inconsistent. six gains are for crop relicts. Four of the gains are records of subspecies where previous recording was only to species level. Thus of the 75 added taxa, up to 42 may be a result of changes in recording policy. The remaining 33 are listed below.

It is difficult to assess which of these are real gains and which have simply been missed earlier. Plantago coronopus (Buck’s-horn Plantain) and Lepidium ruderale (Narrow-leaved Pepperwort) are both plants that have been spreading along main road verges with Cochleria danica (Danish Scurveygrass) and other halophytes - these gains are almost certainly genuine. Prunus x fruticans, the hybrid between Plum and Blackthorn, was not recognised in the county until 1997 but has probably been present for a very long time. The three Bromus grasses were all found in wheat fields where they were probably the result of foreign seed impurities. A striking feature of the gains is the very high proportion that are plants of ruderal or disturbed habitats. The only exceptions are five plants of ponds and riversides, Carex remota (Remote Sedge), Glyceria fluitans (Floating Sweet-grass), Juncus acutiflorus (Sharp-flowered Rush), Ranunculus aquatilis (Common Water-crowfoot) and Veronica anagallis aquatica (Blue Water-speedwell). All the gains are rare or scarce.

Analysing losses brings a different set of questions. As well as asking if the lost plant has just been missed in the current survey, it is important to assess the validity of the old record. This involves judgement about a series of factors. Is the record plausible? Was the recorder known to be reliable? Is there any other confirmation? What features were used in identification at the date of the record and are they now considered diagnostic? Is the locality of the record certain? Making these judgements can take a lot of research and will often involve looking at other records made by the recorder. If he or she is regularly recording a rare species but not a more common confusion species, then an error seems likely. If there is a series of records for a species outside the known range, especially if other experienced recorders are known to have been active in the area, then reliability must be questioned. A springtime record for a taxon requiring ripe fruits for positive identification must be doubted. If the lost plant is in a place outside the area where the recorder normally operated, then grid reference should be checked against other information as far as is possible - if one record from Birmingham is found on the same date as 30 from Malvern from the same recorder then suspicions are raised, especially if the Birmingham record is for a characteristic Malvern plant.

Such an assessment has been carried out for the lost records on the two tetrads. A handful of records have been rejected for various reasons, and it is impossible to rule out further errors, but the current analysis is based on records that there is no good reason to doubt. It is unlikely that more than one or two lost records in each tetrad are based on previous errors of recording. SO94A has 61 lost records that are summarised below. 31 of these records were localised to a six-figure grid reference, and can therefore be specifically looked for. The remaining 30 were only recorded to the tetrad level and it is more difficult to know if they are still present but unseen.

11 losses are from a roadside lay-by at Egdon, all the records being from 1992. This lay-by was taken over by travellers who were finally evicted by the Council who then blocked the lay-by with builders rubble and dumped soil. For a few years afterwards, a strange mix of garden plants, bird-seed aliens and other casuals appeared. By 2003 nearly all of these had been swamped out by coarse grass and scrub. Seven plants characteristic of unimproved grassland were noted on the railway embankment in 1988. These plants have disappeared together with nearly all the grassland as the embankment has been taken over by scrub and trees. It is possible that some of the seven may still linger on, as a few tiny patches of grassland still have a few plants of Viola hirta (Hairy Violet) and Knautia arvensis (Field Scabious). In a flush in an unimproved pasture by the railway two interesting plants were found previously - Eleocharis uniglumis (Slender Spike-rush) in its only hectad site, and Oenanthe fistulosa (Tubular Water-dropwort). The grassland has been improved, the flush has dried up and the plants have gone. This was probably the site where Fragrant Orchid was seen in 1867. Near Mucknell Farm is a reed bed / wet grassland area from which three plants have been lost. They are Apium graveolens (Wild Celery), Equisetum palustre (Marsh Horsetail) and Juncus subnodulosus (Blunt-flowered Rush). These three good native wetland plants have probably gone as the area has dried-out and become more overgrown with rank vegetation. The drying-out may be a natural result of the recent dry summers of the 1990s (and 2003). The other localised plant losses are a mixed bunch. Senecio vulgaris var. hibernicus (Rayed Groundsel) grew in a pub flower-bed from 1987 to about 1992 but then disappeared. Veronica agrestis (Green Field-speedwell) is a scarce arable weed - it may well still exist elsewhere in the tetrad. Ulex europaeus (Gorse) and Daphne laureola (Spurge-laurel) have both disappeared from the verges of the main road near Lowhill. These verges have been subject to considerable alteration over the years. From the same verge was Sinapis alba (White Mustard) - this was probably just a casual. Euphorbia lathyris (Caper Spurge) is a garden escape. Lathyrus nissolia (Grass Vetchling) is very easily over-looked outside its short flowering season and could still exist on the grassy banks of the farm reservoir: if it was grazed by the Canada Geese it would be almost impossible to spot. The final localised plant, Dryopteris dilatata (Broad Buckler-fern), has been lost from the small Lowhill Coppice.

Six un-localised losses are usually woodland plants and may well have gone from the same site - there is very little woodland in the tetrad. Ten are wetland plants, characteristic or open water, pond and river edges, or wet grassland. Four are normally grassland plants, six are probably ruderal and two arable. The Digitalis purpurata (Foxglove) record is probably a garden escape, since this species has no clearly native populations nearby. The Pyrus pryaster (Wild Pear) is probably not a real loss, since this name has in the past been used for wild and cultivated pears - Pyrus communis is present in the tetrad. The distinction between the two pear species is difficult with many intermediate trees in the county. For SO95A the pattern of losses is different from the pattern of gains. Less than half the losses (45%) are plants of ruderal / disturbed habitats. 26% are wetland plants, 18% are grassland plants and 11% are woodland plants. Only 16 of the gains were of native plants or archeophytes (long-established introductions) as against 35 of the losses.

The 68 gains in SO95W included 18 planted trees and shrubs, 15 garden escapes / throw-outs , 4 crop relicts and 3 subspecies previously recorded only at the species level. Thus 40 of the 68 additions may be the result of changes in recording policy. The remaining 28 are shown below:

As with SO95A, it is difficult to know how many of these are real gains. Just over half are plants of ruderal and disturbed habits, a significantly lower proportion that in SO95A. The remainder are more mixed, with six plants of wetland, three of grassland and three of woodland. Woodland is rare in the tetrad, and two of the three woodland plants were found in Abberton churchyard. There are no obvious gains like the road verge Scurvygrass in SO95A, and it is even possible that all the additions were present during earlier surveys but were not spotted. On the other hand, many of the sites for the new records were well covered in at least one earlier survey. Abberton Church was surveyed at least three times in the 1980s, Whitsun Brook was covered in a 1982 Trust river survey, Baynhall Meadow was surveyed in 1981 and 1991, and a WFP field meeting spent a full day in the tetrad as part of the 1987/88 survey.

The 50 losses from SO95W are shown below:

Unfortunately, only 20 of the 50 losses are localised to a specific spot within the tetrad. This makes it hard to target searches for some of the lost taxa, so a few may have been missed in 2003. Baynhall Meadow is an SSSI grassland site which was not adequately covered in 2003; a further visit in 2004 may well turn up some of the six losses from here. Three of the losses may be misinterpretation or errors. It was assumed that the 1988 record for Lamiastum galeobdolon (Yellow Archangel) was for the native subspecies ssp. montanum. A thriving colony of the introduced ssp. argentatum has been present between 1991 and 2003 and the earlier record probably refers to this. Papaver dubium (Long-headed Poppy) was assumed to be ssp. dubium but subsequent records show the presence of ssp. lecoqii. The record for Crataegus laevigata (Midland Hawthorn) may be correct but recent recording has only found the hybrid with common Hawthorn, which can look very similar. Despite these provisos most of the losses will be genuine, since most known or likely sites have been covered well.

There has been less change in habitat than with SO95A. The area of grassland near Abberton Hall where Helictrotrichon pubescens (Downy Oat-grass) and Lathyrus nissolia (Grass Vetchling) were recorded has become very overgrown with coarse vegetation although an addition from the remnant area was Festuca ovina (Sheep’s Fescue) on an old ant-hill. The main change has been to the pool by Baynhall Meadow. In 1987/88 this was a fairly small field pond in pasture. It is now a large fishing pool following substantial earth moving and the flora has changed with both gains and losses. The losses can be analysed by known or presumed habitat. If the three likely errors are omitted we find 13 losses from grassland habitats, 16 from water and wetlands, 15 from ruderal or arable areas and 3 from woodland. Some of the unlocalised records are difficult to classify to habitat: for example Viola riviniana (Common Dog-violet) has been treated as a woodland plant, although with this predominantly agricultural tetrad it may be more likely that it was a ruderal garden escape.

The overall analysis of gains and losses is shown below:

What can be deduced from this? Apart from the more critical recording of sub-species, the net gains are all from habitats heavily influenced by man. The planted trees and shrubs, the garden escapes and the crop relicts are very directly the result of human activity. The ruderal and arable plants show a considerable turnover of taxa with moderate net increases. This turnover is to be expected. Disturbed ground is available to good colonisers with efficient means of dispersal, so such plants can come in from a wide area. The disturbed habitats are normally short lived, however, with either slower colonisation by more competitive, usually perennial, species or recurrent disturbance as with arable fields. The net increase in these habitats may be exaggerated by more intensive recording effort in 2003, but is probably real.

Semi-natural habitats have fared much worse. Grassland, wetland and to a lesser extent woodland all show net decreases. Little can be deduced from the woodland changes, since the two tetrads have no significant areas of semi-natural woodland. The grassland and wetland losses are significant and worrying. There is a clear reduction in grassland floral biodiversity in both tetrads. Unimproved grassland has been rare throughout the period since 1970, and of the little that was present at the beginning of the period most has been destroyed. Only the SSSI Baynhall Meadow remains as a good example of this habitat: agricultural intensification and scrub encroachment have destroyed nearly all the rest. This pattern is repeated throughout Worcestershire and soon our native grassland species will be virtually restricted to small and scattered specially protected sites. The picture with wetlands is more complex. True aquatics tend to be either good colonisers or plants that can survive long periods of unfavourable conditions in the seed bank. As with ruderals, a high turnover of taxa is unsurprising. The drying out of ponds and the growth of over-shading vegetation on our brooks and pools has probably made conditions unsuitable for many of these true aquatics and where sites are still in an appropriate condition, the original plants have often been replaced with others. Increased eutrophication may also be a factor in the decline but the data are insufficient to demonstrate this conclusively. The wetland taxa of damp grassland have suffered particularly badly with no gains to offset the losses. Agricultural intensification and drainage, and the recent run of dry summers have combined to make the plants of damp grassland among our most threatened across the county. There is no sign that this tendency is likely to reverse.

In summary, an overall increase in floral biodiversity, which is probably genuine, is masking a significant decline in our native species of “good” habitats. These native plant losses tend to be poor colonisers that are unlikely to return even if the habitats can be recreated or recovered. There are few local sites from which re-colonisation can occur. The loss of Eleocharis uniglumis (Slender Spike-rush) from the flush in grassland by the railway may well be extinction at the hectad level.

Lessons can be learned about the design of surveys to pick up floral change. Only by standardising the route taken and the time spent at specific seasons can repeated surveys of a single tetrad hope to demonstrate change conclusively. Even with such standardisation, problems will arise with the variable skills of individual recorders (or even the same recorder over time). Full localisation of all taxa rather than just those thought to be less common will make it easier to be sure that losses are genuine and to explain why such losses have occurred. The rarity of many plants at the tetrad scale makes it likely that some of the most important and interesting plants will be missed without fairly exact knowledge of where they grew.