The Effect of Ecological Restoration on the Structure and Function of Soil Microbial Communities in Cranberry Bogs

In the summer of 2022, the Ballantine and Andras research groups from Mount Holyoke College collected 255 soil samples from 26 localities across Eastern Massachusetts as part of a large and ongoing cranberry restoration monitoring project funded by the Massachusetts Department of Fish and Game, Division of Ecological Restoration (DER). To facilitate a comparative analysis of wetland soil development, our sampling localities were chosen to encompass four different site types: sites that were in active cranberry agriculture, sites that were retired from cranberry agriculture, former cranberry farms that have undergone ecological restoration, and natural reference peat bogs. These localities were geo-referenced and coordinated across all working groups involved in this study in order to facilitate comparative analysis across many types of data including attributes of soil biogeochemistry, as well as communities of plants, invertebrates, amphibians, reptiles, birds, and fish.

Our initial analysis, summarized in our 2023 DER Soils Monitoring Report, focused on physicochemical soil attributes (e.g. pH, soil moisture, bulk density, soil organic matter, cation exchange capacity, etc.). This year, we received funding from DER to complement our prior results by analyzing the soil microbial communities from the same samples collected in 2022. To do so, we employed 16S amplicon sequencing to survey all species of resident microbes and then used statistical ecological analyses to assess and compare the communities of different soil types (restored, farmed, retired, natural).
 
Our research group has conducted previous studies indicating that ecological restoration promotes shifts in soil microbial communities and associated biogeochemical functions, making them more similar to undisturbed reference peat wetlands (Andras et al. 2020; Rubin et al. 2022). However, we did not observe such definitive patterns in the present monitoring dataset. Figure 1 (below) illustrates the broad taxonomic composition of each site at the phylum level, with sites grouped by restoration status. And Figure 2 (below) depicts a three-dimensional principle-coordinates (PCoA) plot of the microbial communities of each soil sample, with the proximity of points on the plot indicating their degree of similarity. It is evident in both figures that the undisturbed reference sites are fairly distinct from the rest – they have a significantly different pattern of taxonomic composition (e.g. on average, more Acidobacteria and fewer Proteobacteria and Actinobacteria than other site types – Figure 1), and they compose a tight and mostly exclusive cluster on the PCoA similarity plot (orange points - Figure 2). In addition, active farms form a diffuse cluster in the similarity plot that is somewhat distinct from most other samples. However, samples from retired and restored sites are broadly dispersed on the PCoA similarity plot. With few exceptions, they do not appear to be shifted toward the natural reference sites, and their composition was not significantly different from farmed or retired sites.
 
This raises the question of why the present monitoring dataset has not reproduced the patterns observed in multiple previous studies. There are several possible explanations for this. First, the reference samples that were used in the present data set were different from those used in previous studies, and their sediment was noticeably distinct. In our prior studies, the chosen reference sites were kettle-hole wetlands where the top sediment layer was composed of peat mixed with heavy, dark, organic muck. In the present dataset, we selected as our reference sites Atlantic white cedar bogs that were immediately adjacent to active cranberry farms. These sites were chosen based on consultation with other working groups and the presumption that the adjacent cranberry farms were constructed on similar substrate. However, these reference samples were composed almost entirely of undecomposed sphagnum with very little decayed muck at all. It is possible that this substrate fostered a microbial community that was distinct not only from the other restored, retired, and farmed soils, but also from the reference soils of previous studies. If the microbial community of the newly chosen reference sites were sufficiently distinct, it could obscure more subtle shifts in the communities of restored soils.  A metanalysis of these studies could resolve this question. Nonetheless, this does raise an important question that is a topic of ongoing discussion among the Living Observatory working groups: What are appropriate reference sites for our monitoring efforts? This question warrants further discussion and careful consideration.
 
A second potential explanation for the disparity between past and present studies is sampling time. The current data set was collected in early May of 2022, while the previous study samples were collected in late June and July. It is possible that seasonal differences in temperature, moisture, and/or vegetative activity could obscure more subtle patterns of similarity and difference that could become more evident later in the season.
 
A third possible explanation for the differences among studies is sampling effort. The depth of genetic sequencing was moderately greater for the two previous studies than it was for the present study, meaning that the microbial community was surveyed more intensively. It is possible that the deeper sampling of our previous studies recovered more rare taxa that were shared between restored and reference samples, thereby elucidating more subtle similarities.
 
Finally, it is important to note that all of the six restored cranberry bogs available for study are still quite young, with an average age of less than four years at the time of sampling. Given the preponderance of prior evidence on hydrology, plants, soils, and microbes, it is certain that ecological restoration of cranberry farms does put these sites on a developmental trajectory toward natural reference sites. The primary remaining questions for future monitoring efforts are what appropriate reference sites should be, how quickly restoration can meaningfully alter relevant soil characteristics, how pre-existing site attributes affect restoration trajectories, and what techniques might be employed to more efficiently and effectively promote desirable changes.