34
Kelp Studies
Satellite imagery has also proved useful for mapping
kelp distributions. High resolution satellite imagery
acquired by SPOT and WorldView has been validated by
kayak surveys. Results show that the technique has good
accuracy when image collection conditions are suitable.
When tides are low and kelp beds are dense or mostly
on the surface, kelp can be detected with high accuracy
(Figure 6).
However, conditions in historical imagery including tide,
date and surface conditions are not optimized for remote
sensing of kelp and thus kelp detected in these images
does not necessarily represent the true maximum extent
of kelp present in that year. For this reason, researchers
using satellite images will denote kelp presence or
absence in defined regions rather than trying to quanti-
tatively assess the extent of kelp. Further refinements to
the methodology for analyzing temporal data are being
explored to improve the ability to validate historical
classifications.
SIGNIFICANCE AND NEXT STEPS
Next steps for the phytoplankton studies will be the
linkage of further elements up the food chain to salmon
— namely, addition of information on zooplankton,
focussing on those species most important in salmon
diets. Results will provide insight into how variations in
both phytoplankton and zooplankton may impact growth
and survival of juvenile salmon in the Strait of Georgia.
Figure 6. Results of satellite classification (red) and ground-truthing polygons from kayaks (green lines) have 81% overall accuracy.
Note some inaccuracies are due to error in the ground-truthing data, which was collected by kayak and GPS. Figure from Sarah Schroeder.
Groundtruthing: Kayak Mapping Polygon Classification from Satellite Data
