Gentle approach was needed
Grontmij’s technical manager, Harriet Carlyle, and the consultancy’s former head of ecology, Simon James, explain how potential hydro-ecological impacts of a utility’s water abstractions were assessed using modelling
The ecological value of these sites is due to the presence of wetland plant communities and associated fauna, which are dependent on the specific physical and chemical conditions associated with shallow groundwater and surface water.
Engineering consultancy Grontmij, part of the @one Alliance, is investigating these concerns on behalf of AWS under the AMP4 Water Resources Environment Programme. In particular, the potential impact of each abstraction on the groundwater and surface water-dependent ecosystems are being assessed.
The protected conservation sites range from river habitats in Norfolk and groundwater-fed wetlands in Cambridgeshire to clay pits in North Lincolnshire. At the start of the project, the sites to be investigated were at various stages in the Environment Agency's Habitats Directive Review of Consents (RoC) process.
The majority of the sites were at stage three, and so an appropriate assessment was carried out to establish whether the permissions - including licensed abstractions - were having an adverse effect. A number of the sites were at stage four, in which permissions were being affirmed, modified or revoked.
A handful of sites that were classified as SSSIs, and therefore not part of the RoC process, were also included for investigation. The Environment Agency is using the findings of the @one Alliance investigations to assist in producing Site Option Plans and Site Action Plans.
For sites at stage three of the RoC process, an initial desk study was undertaken prior to any site investigations to gain an understanding of the hydrogeology, hydrology and ecology of each site as well as the eco-hydrological requirements of the identified wetland communities and the potential impact of abstractions. In addition, Environment Agency regional groundwater model runs were commissioned for a number of sites to assess the modelled impact of Anglian Water Services abstractions on surface water and groundwater levels. Site investigations designed to fill data gaps included the construction of observation boreholes and dipwells, long-term hydrometric monitoring, spot flow gauging, pumping (signal) tests, topographic and bathymetric surveys, hydrochemical surveys, geophysical surveys, and river habitats and botanical surveys.
For all SACs and SSSIs, a Hydrological Impact Assessment was produced, bringing together the results of all studies to inform on the impact of AWS abstractions on site. Challenges included:
- Understanding the - often complex - hydrogeology of the wetland sites
- Establishing the ability of the regional groundwater models to simulate impacts on a local scale
- Assessing impacts as a result of other activities, such as nitrate loading, discharges, land drainage and land management
- Agreeing what constitutes an acceptable target - maximum change in groundwater or surface water level or flow so as not to cause an adverse impact on the ecosystem
- Establishing baseline data for monitoring any future change
The direction of any change in species distribution and/or abundance, and the particular species affected, helps establish whether the changes are attributable to changes in groundwater or surface water regimes associated with abstraction. Traditional methods of mapping plant distribution, such as the National Vegetation Classification (NVC) system, rely on subjective judgments by the surveyor.
Therefore any difference over time could quite easily be attributable to a change in the survey team. In addition, NVC does not provide details of individual species distributions, so is unsuitable for monitoring changes in those less abundant or rare species that may move, decrease or disappear due to changes in groundwater availability or chemical composition.
More robust surveys may be carried out along transects (Wheeler-Shaw method), which can be effective when the location of any likely changes in plant distribution are known, but cannot account for range shifts in unexpected parts of the site. Geostatistical vegetation mapping (GVM) is the application of a geostatistical technique called kriging to statistically predict the distribution and abundance of plant species and communities found on a site using GPS- referenced quadrat data.
Kriging techniques are used to produce the GIS probability maps, effectively interpolating the missing data between the known data points. While geostatistical techniques were developed more than 40 years ago to predict the distribution and abundance of precious metal ores, these same techniques have only relatively recently been introduced to the natural sciences.
Grontmij worked in collaboration with staff from Cranfield University and Professor Isobel Clark to refine the geostatistical methodology, working in conjunction with Natural England, the Environment Agency and AWS, to apply it to the hydroecological studies described above.
Grontmij also mapped the Ellenberg values for distribution of plant species using their requirements for water, nutrients, calcium ions, light and pH. This allowed probability maps to be produced that showed the distribution and relative quantities of these environmental variables. It is any future change in these probability maps, based on shifts in plant species, that will allow assessment to be made as to the probable cause, such as the reduction in available calcareous groundwater or an increase in nutrients. The ecological data is supplemented with data on hydrochemistry, hydrogeology and land management practice to determine if changes in these variables could also be significant co-contributors to any observed changes in vegetation patterns.
A further advantage of GVM is that future assessments should help pinpoint whether changes in the groundwater regime are responsible for any changes to the ecology of a site, as opposed to, or in tandem with, other factors such as nutrient infiltration from adjacent farmland. In contrast, other mapping methods may not allow a sufficiently robust review of data between repeat surveys, which as a consequence can make it difficult to identify trends.
Further advantages of GVM include:
- Repeatability - the field survey simply records if a plant is present at a point (quadrat). There is no arbitrary and variable drawing of community boundaries. Also, the kriging technique produces a continuous surface between point data, but these survey points do not need to be precisely replicated during subsequent surveys. Following the same approximate grid system during each field survey produces kriged surfaces which can be compared on a statistically sound basis
- Accuracy - he quadrat point data positions are recorded to sub-metre accuracy, and the kriged surfaces use the results of many-point data to determine with confidence the probability of species occurrence. Also, as the surveys are carried out for individual plant species, the results are highly sensitive to any changes in distribution
- Unbiased - as the data is collected using an unbiased sampling methodology, results may be compared over time using statistics, as in a standard ecological study. This means that confidence intervals can be assigned to any distribution changes, essential for objective assessment of results