Detailed Scientific Methods
INDICATOR DEVELOPMENT
1. Measurement Type
There are several approaches, that are presently used to monitor kelp, falling into two primary classes: (1) in-situ surveys conducted by divers and (2) overhead surveys that measure the extent of emergent kelp canopy on the ocean surface. The Kelp Indicator relies on overhead surveys to report on the status and trends of kelp, as these surveys, collectively, offer both greater spatial coverage and temporal fullness and duration than that of underwater surveys. For example, measurements of kelp canopy at the surface are routinely collected along the U.S. West Coast from both aerial overflights and multiple satellite platforms. However, because overhead surveys only measure what is at the surface, this Kelp Indicator does not include understory species that don’t reach the sea surface. This is an assessment of the two primary canopy-forming kelp species along the U.S. West Coast, bull kelp and giant kelp (Nereocystis luetkeana and Macrocystis pyrifera). The metric for kelp canopy extent can differ among programs. For example, surveys can differ in whether they measure total kelp canopy extent, kelp canopy coverage, or kelp canopy biomass. This assessment is of kelp canopy coverage expressed as an area of sea surface occupied by kelp.
2. Datasets
The Kelp Indicator relies on SBC LTER: Time series of quarterly NetCDF files of kelp biomass in the canopy from Landsat 5, 7, and 8 since 1984 (ongoing). This dataset is published quarterly and includes a time series of canopy area of giant kelp and bull kelp derived from Landsat 5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM+), Landsat 8 Operational Land Imager (OLI), and Landsat 9 Operational Land Imager 2 satellite imagery, along with relevant metadata. The different Landsat sensors have been calibrated to each other using simulated Landsat data derived from hyperspectral imagery. Missing data due to the ETM+ scan line corrector error were filled using a synchrony-based gap filling method. Canopy area (m) data are given for individual 30 x 30 meter pixels for all coastal areas (including offshore islands) from California, Oregon, and the outer coast of Washington. This data product does not include Washington’s inner coastal areas.
To account for differences in both the seasonality and perennial nature of both canopy-forming species, the annual maximums from the quarterly data are used, producing a spatially resolved annual time series of kelp canopy coverage since 1984.
While the indicator relies on Landsat-derived surveys of kelp canopy coverage, comparison with other overhead surveys have been conducted. Results are forthcoming
3. Unit of Assessment
To assess kelp, the U.S. West Coast is divided into semi-regularly spaced coastal segments of ~5-10 km. In total there are 360 coastal segments, of which 233 have historically contained kelp, referred to as kelp-containing coastal segments. We define a coastal segment as historically containing kelp based on detection of at least 10 pixels in the Landsat time series. This is equivalent to 0.009 km2 of a historically detected spatial extent. Fifty of the coastal segments have no detection of kelp (zero pixels), and six fall into this latter category of minimal detection.
The California coastline was segmented into ~10-km segments along the mainland and Northern and Southern Channel Islands off southern California. Coastal segments extend from the coastline to the state water boundary, 3 nautical miles from the shore. California has 196 coastal segments, 146 of which have historically contained kelp.
The Oregon coastline was segmented into ~5-km segments, adopted from polygons generated by Oregon’s Department of Fish and Wildlife (courtesy Scott Marion). There are a total of 82 coastal segments, with the first 81 as regularly-space segments. The northernmost segment is X km in length, and extends a portion of the coastline that historically contains no canopy-forming kelp. Of these, 27 segments have historically contained kelp.
The Washington coastline was segmented into ~5-10 km segments, adopted from polygons generate by Washingtons Department of Natural Resources (link). We assess the outer coast and western section of the Strait of Juan de Fuca. Here there are 82 coastal segments, 60 of which have historically contained kelp. This assessment does not include the eastern straits or Puget Sound of Washington because the coarse resolution of Landsat impedes its ability to accurately detect kelp in this region. However, these parts of Washington’s inner coast are included in Puget Sounds Vital Signs, and their findings are integrated here.
Some coastal segments, while historically containing kelp, its presence is too ephemeral to make a status and trends assessment based on historical expectation (low signal-to-noise ratios). There are 53 of these segments; 18 in California, 11 in Oregon, 6 in Washington. There are an additional 18 segments in Washington for which Landsat data is unavailable. The coastal segments can be viewed in our Marine Planner visualization tool HERE.
4. Status
For the Indicator, the status is based on the most recent year for which kelp canopy extent is available. The extent of kelp within a coastal segment is standardized relative to a fixed historical baseline and expressed as a percentile. Such an approach gives equal weighting to all kelp-containing coastal segments. The status for the entire U.S. West Coast is categorized as Far Below, Below, Meeting, Above, or Far Above Expectation. The Expectation categories are defined from the historical distribution of median coast-wide coverage (Figure). The current year’s median coverage is assessed relative to this historical distribution and represents the needle on the gauge.
5. Trend
Two trends are assessed: (1) linear trend for the full historical period for which data is available (1984-current year), and (2) linear trend for the time period since 2017. This latter period is to assess trends in kelp since the 2014-15 marine heatwave, ‘the blob,’ which resulted in significant kelp losses along many parts of the West Coast, and to which recovery rates have differed. Based on the significance and magnitude of these two trends, kelp-containing coastal segments are categorized as Degrading, Slightly Degrading, Staying the Same, Improving, and Slightly Improving.
INDICATOR REVIEW
Please pardon our dust. This section is in progress.
INDICATOR CONFIDENCE
The Data Quality of the Kelp Indicator is an 8 (out of 10). The Landsat‐derived dataset of kelp canopy extent offers 40 years of historic perspective, it extends the full U.S. West Coast, and provides near‐real time data. However, Landsat is one of the oldest, and therefore coarsest, satellite products. As such, it is challenged to detect patchy kelp and kelp near the coast. Currently, 8% of canopy‐forming kelp along the coast cannot be assessed with this survey. Newer satellites provide improved overhead surveys, but it will take several years of co‐occurring data to create a translation from the historic to the more recent data.
CALCULATION METHODOLOGY
Please pardon our dust. This section is in progress.
Placeholder: Detailed Methodology and equations (webpage or pdf)
LOOKING AHEAD
Please pardon our dust. This section is in progress.