Modelling kelp forest distribution and biomass along temperate rocky coastlines

Daniel Gorman ∙ Touria Bajjouk ∙ Jacques Populus ∙ Mickaël Vasquez ∙ Axel Ehrhold

Département Dynamiques de l'Environnement Côtier, Institut français de recherche pour l'exploitation de la mer, BP 70 Plouzané, France 29280

Corresponding author:

Supplementary Material

Here we provide additional information on the methodology used to derive thirteen abiotic environmental variables used to model kelp forest distribution, proportional covers and biomass. The variables developed can be broadly grouped into the categories; water column characteristics, seafloor characteristics, light availability and wave exposure. Datasets were pre-processed in ArcGIS (raster grid format) so that they were all in the same geographic reference system (French Lambert 93).

Water column characteristics

We obtained data summarising mean sea surface temperate, salinity and nitrate concentrations for the English Channel region from the French forecasting system PREVIMER This data archive stores forecasts calculated each hour in 4 km resolution grids. Datasets were averaged across 3 consecutive years (2008-2010) and reprocessed to give a raster grid layer with a 25 m2 horizontal resolution.

Seafloor characteristics

A Digital Terrain Model (DTM) was developed for the Bay of Morlaix by merging data obtained from Light Detection and Ranging surveys and bathymetric data supplemented by DGPS positioned acoustic soundings that were interpolated using kriging software (ISATIS®, Geovariances, Cedex, France). The DTM was used to produce a bathymetric raster layer (25 m2 horizontal resolution). Benthic Position Index was derived from the DTM using the methods of Weiss (2001) using the geo-processing tool ‘Benthic Terrain Modeler’ Seafloor slope was derived from the DTM using the Benthic Terrain Modeller extension to ArcGIS spatial analyst tool. Rugosity units were derived from the DTM using the algorithm of Sappington et al. (2007) and ranged from 0 (flat) to 1 (most rugged) (25 m2resolution). Within polygons of subtidal rock sediment proximity was defined as the linear distance from the closest area of unconsolidated sediments calculated using the interpreted rock polygon layer. (25 m2resolution).

Light availability

We investigated the effect of light availability across a hierarchy of scales; (1) broad-scale (Atlantic and English Channel waters) estimates of surface light attenuation coefficient (KdPAR) and depth-attenuated light fraction (ZEU), and (2) a local-scale (25 m2) index termed ‘insolation’ which provides a shaded relief representation of the DTM that can influence benthic communities at the scale of meters. Temporally integrated measures of KdPAR and ZEU were obtained by averaging MERIS reflectance data (see Saulquin et al. 2011) for coastal European waters over five consecutive years (2005 – 2009). Insolation was calculated by quantifying the total duration of solar exposure (unit hours/year) for each 25 m² pixel using the ‘Potential Incoming Solar Radiation’ tool provided by SAGA GIS Software (Böhner et al. 1997).

Wave exposure

To quantify the degree of wave exposure experienced by subtidal kelp forest communities (i.e., the combination of drag force and substrate orientation) we created indices describing mean significant wave height (SWH) and deviation from predominant swell direction (WDEV). Significant wave height was calculated using the long-term average hind cast predictions (2006-2010) provided by IOWAGA (Integrated Ocean Waves for Geophysical and other Applications) Project, which is based on WAVEWATCH III modelling framework (250 m2 resolution). Wave direction was calculated using the DTM layer to produce an aspect layer from which the angler deviation from average annual wave direction (provided by IOWAGA, see above) was calculated using GIS tools (250 m2 resolution). This index ranged from 0º (i.e., directly facing oncoming swell) to 180º (i.e., facing leeward).

Table S1.Summary of the thirteen environmental variables used to model the distribution and structure of Laminarian kelp forests in the Bay of Morlaix.

Predictor / Units
Seafloor characteristics
Water depth † / m
Benthic Position Index † / BPI units
Rugosity † / Rugosity units
Distance from sediments † / 1/m
Slope † / ̊from horizontal
Light availability
Photic depth ‡ / Fraction (%)
Insolation ‡ / h year-1
Surface light attenuation coefficient ‡ / KdPAR
Exposure gradients
Significant wave height נ / m
Wave deviation ◊† / angler°
Water column characteristics
Nitrate concentration ◊ / µmol L-1
Salinity ◊ / psu
Sea surface temperature ◊ / °C

Data sources, † = IFREMER, ◊ = PREVIMER: a project managed by IFREMER and supported by funds from the European Union, with SHOM and Météo-France as principle partners, ‡ = EUSeaMap: mapping European seabed habitats (

References

Böhner J, Köthe R, Trachinow C (1997) Weiterentwicklung der automatischen Reliefanalyse auf der Basis von digitalen Geländemodellen. Göttinger Geographische Arbeiten 100: 3-21

Sappington JM, Longshore KM, Thompson DB (2007) Quantifying landscape ruggedness for animal habitat analysis: A case study using bighorn sheep in the Mojave Desert. Journal of Wildlife Management 71: 1419-1426 doi 10.2193/2005-723

Weiss A (2001) Topographic position and landforms analysis. ESRI User Conference, San Diego, CA.

1