Soil eDNA helps Jacobs survey biodiversity for road project

Jacobs were able to visualise the difference in soil biodiversity between grassland and woodland habitats at their site, enabling them to characterise the ‘gap’ that needs to be bridged through habitat creation.

They now have a solid baseline against which to monitor as the project progresses. It is anticipated new samples will be collected in subsequent years and added to this dataset to show how communities are changing over time and relative to the baseline and control sites.

Jacobs can now objectively compare the soil communities of the different woodland types to inform the soil translocation mitigation strategy and establish which soils are best suited to translocation.

Figure 1. An ordination plot for soil faunal communities. Each point represents one sample and is coloured by habitat. Points that are closer together have more species in common than points that are further apart.

Soil eDNA provided a dataset of almost 3,000 species-level taxa.

There were clear variations in the communities associated with the different types of habitat. In particular, grassland soil communities differed significantly from those in the woodland that would likely be affected by the road scheme (Figure 1). The grassland habitat also had lower fungal and faunal richness (number of taxa) compared to the woodlands.

The soil faunal and fungal datasets independently showed very similar patterns of variation among the different habitats, which adds confidence to interpretations. Bacteria were slightly less informative for differentiating habitats but still varied significantly between the woodland and grassland habitats.

There was no significant difference between the soil communities of the ancient birch woodland and the non-ancient birch woodland. Additional sampling would be required to strengthen this finding, but it may have implications for the mitigation strategy in terms of the range of woodland soils that could be used to assist the successful establishment of new woodland compensatory planting sites.

We also carried out analysis of soil fauna that had been extracted from large 100cm3 soil cores using heat sources, as is typical for morphological analysis. This yielded only around 10% the number of species detected by direct analysis of the soil samples, with a strong bias towards arthropo (especially beetles, mites and springtails). Major groups such as nematodes were entirely missing and the power of this data to characterise and differentiate habitats was much lower. Furthermore, direct metabarcoding of the soil samples carried less risk of samples being lost through transporting between laboratories and overall was cheaper than the two-step method of pre-extraction then DNA metabarcoding.

This survey was able to set the baseline for habitat creation. As grassland areas are planted to create woodland, one of the success measures will be the extent to which the soil communities become more like those in the woodland baseline, indicating the establishment of a truly representative woodland habitat. This could be accelerated by mitigation strategies such as translocating woodland soils to grassland areas.