Automated satellite remote sensing of giant kelp at the Falkland Islands (Islas Malvinas)

Abstract

Giant kelp populations that support productive and diverse coastal ecosystems at temperate and subpolar latitudes of both hemispheres are vulnerable to changing climate conditions as well as direct human impacts. Observations of giant kelp forests are spatially and temporally uneven, with disproportionate coverage in the northern hemisphere, despite the size and comparable density of southern hemisphere kelp forests. Satellite imagery enables the mapping of existing and historical giant kelp populations in understudied regions, but automating the detection of giant kelp using satellite imagery requires approaches that are robust to the optical complexity of the shallow, nearshore environment. We present and compare two approaches for automating the detection of giant kelp in satellite datasets: one based on crowd sourcing of satellite imagery classifications and another based on a decision tree paired with a spectral unmixing algorithm (automated using Google Earth Engine). Both approaches are applied to satellite imagery (Landsat) of the Falkland Islands or Islas Malvinas (FLK), an archipelago in the southern Atlantic Ocean that supports expansive giant kelp ecosystems. The performance of each method is evaluated by comparing the automated classifications with a subset of expert-annotated imagery (8 images spanning the majority of our continuous timeseries, cumulatively covering over 2,700 km of coastline, and including all relevant sensors). Using the remote sensing approaches evaluated herein, we present the first continuous timeseries of giant kelp observations in the FLK region using Landsat imagery spanning over three decades. We do not detect evidence of long-term change in the FLK region, although we observe a recent decline in total canopy area from 2017–2021. Using a nitrate model based on nearby ocean state measurements obtained from ships and incorporating satellite sea surface temperature products, we find that the area of giant kelp forests in the FLK region is positively correlated with the nitrate content observed during the prior year. Our results indicate that giant kelp classifications using citizen science are approximately consistent with classifications based on a state-of-the-art automated spectral approach. Despite differences in accuracy and sensitivity, both approaches find high interannual variability that impedes the detection of potential long-term changes in giant kelp canopy area, although recent canopy area declines are notable and should continue to be monitored carefully.

Kelpwatch: A new visualization and analysis tool to explore kelp canopy dynamics reveals variable response to and recovery from marine heatwaves

“…Remote sensing is a powerful tool for monitoring canopy forming kelps such as bull kelp and giant kelp, and recent increases in the availability of airborne and spaceborne imagery is enabling regular monitoring across multiple space and time scales [5,14,22]. For example, inexpensive small unoccupied aerial systems (UAS) can provide very high-resolution monitoring of canopy extent at local scales [23,24], constellations of CubeSats can provide high resolution data on regional scales [25], while moderate resolution satellites can be used to map kelp canopy dynamics at global scales [26,27]. The Landsat satellite program is particularly valuable for kelp monitoring, as it provides imagery with continuous global coverage at a 30 m resolution since 1984.…”

“…Landsat imagery has been used to map both giant kelp and bull kelp canopy density and extent [14,18,28–30] and kelp abundance [18,20] on seasonal time scales from 1984 to present for the west coast of the United States and Baja California, Mexico [14,18,28–30] and other regions of the world [27,31,32]. One of the most valuable aspects of this dataset is its extensive spatial and temporal coverage, especially for distinguishing the impacts of climate change on kelp populations from other sources of variability [5].…”

Kelp Patch-Specific Characteristics Limit Detection Capability of Rapid Survey Method for Determining Canopy Biomass Using an Unmanned Aerial Vehicle

“…Estimation of kelp canopy area has rapidly advanced, leading to regional-scale (e.g Bell et al, 2020b) to global-scale (e.g. Mora-Soto et al, 2020) estimates, and assessments of kelp canopy area in very remote locations (Friedlander et al, 2020;Houskeeper et al, 2022). Despite these advances, it is desirable in some cases to move beyond canopy area to assess biomass, and then primary productivity, as that provides information on kelp bed health, carbon sequestration, and viability.…”

“…To date, a range of platforms with varying spatial resolutions have been applied to kelp mapping efforts (Schroeder et al, 2019). Multispectral USGS Landsat imagery (30 m spatial resolution) has been widely used because of the large temporal and spatial scales at which data are freely available (Cavanaugh et al, 2010;Cavanaugh et al, 2011;Bell et al, 2015;Young et al, 2015;Beas-Luna et al, 2020;Bell et al, 2020b;Friedlander et al, 2020;McPherson et al, 2021;Houskeeper et al, 2022). Studies comparing the suite of sensors (Landsat TM, ETM+, and OLI) to higher spatial resolution aerial imagery (e.g.…”

Artificial intelligence convolutional neural networks map giant kelp forests from satellite imagery

“…Pseudo-RGB composites were generated by selecting the near-infrared (760 to 900 nm), the red and the green bands (Fig. 2 ), in line with recent studies published in the scope of remote sensing of kelp forests 15 , 42 – 44 . While the near-infrared band allows generating images of high contrast, considering the high reflectance of kelp canopies and the high absorption of water masses at this wavelength 19 , 45 , the additional bands (red and green) provide informative parameters to discriminate surface cover type and, for aquatic surfaces, particle content 46 .…”