Consensus forecasting of intertidal seagrass habitat in the Wadden Sea
Folmer, E.O.; van Beusekom, J.E.E.; Dolch, T; Gräwe, U.; van Katwijk, M.M.; Kolbe, K.; Philippart, C.J.M. (2016). Consensus forecasting of intertidal seagrass habitat in the Wadden Sea. J. Appl. Ecol. 53: 1800–1813. https://dx.doi.org/10.1111/1365-2664.12681
Additional data:
In: Journal of Applied Ecology. British Ecological Society: Oxford. ISSN 0021-8901; e-ISSN 1365-2664, more
|   |
| Keywords |
Zostera subg. Zostera marina Linnaeus, 1753 [WoRMS]; Zostera noltei Hornemann, 1832 [WoRMS]
|
| Author keywords |
Boyce index; ensemble forecasting; habitat distribution model; habitat management; machine learning; model transferability; seagrass; Wadden Sea; Zostera marina; Zostera noltii |
| Authors | | Top |
- Folmer, E.O., more
- van Beusekom, J.E.E., more
- Dolch, T
- Gräwe, U.
|
- van Katwijk, M.M.
- Kolbe, K.
- Philippart, C.J.M., more
|
|
| Abstract |
1. After the dramatic eutrophication-induced decline of intertidal seagrasses in the 1970s, theWadden Sea has shown diverging developments. In the northern Wadden Sea, seagrass bedshave expanded and become denser, while in the southern Wadden Sea, only small beds withlow shoot densities are found. A lack of documentation of historical distributions hampersconservation management. Yet, the recovery in the northern Wadden Sea provides opportunityto construct robust habitat suitability models to support management.2. We tuned habitat distribution models based on 17 years of seagrass surveys in the northernWadden Sea and high-resolution hydrodynamics and geomorphology for the entire WaddenSea using five machine learning approaches. To obtain geographically transferablemodels, hyperparameters were tuned on the basis of prediction accuracy assessed by non-random,spatial cross-validation. The spatial cross-validation methodology was combined with aconsensus modelling approach.3. The predicted suitability scores correlated amongst each other and with the hold-out observationsin the training area indicating that the models converged and were transferable acrossspace. Prediction accuracy was improved by averaging the predictions of the best models.4. We graphically examined the relationship between the consensus suitability score andindependent presence-only data from outside the training area using the area-adjusted seagrassfrequency per suitability class (continuous Boyce index). The Boyce index was positivelycorrelated with the suitability score indicating the adequacy of the prediction methodology.5. We used the plot of the continuous Boyce index against habitat suitability score to demarcatethree habitat classes – unsuitable, marginal and suitable – for the entire international WaddenSea. This information is valuable for habitat conservation and restoration management.6. Divergence between predicted suitability and actual distributions from the recent past indicatesthat unaccounted factors limit seagrass development in the southern Wadden Sea.7. Synthesis and applications. Our methodology and data enabled us to produce a robust andvalidated consensus habitat suitability model. We identified highly suitable areas where intertidalseagrass meadows may establish and persist. Our work provides scientific underpinningfor effective conservation planning in a dynamic landscape and sets monitoring priorities. |
|
