GCRMN tier data model and standards
Context
For the analysis of IUCN RLE Criteria, key data sources are required to be standardised and in a similar format. This helps subsequent steps in ensuring specific data columns can be identified and summarised for deriving the criteria data objects, evaluation of status, and visualisations.
The Spatial and Physical data for the evaluation of Criterion B and Criterion C (i.e. Restricted geographic distribution and Environmental degradation, respectively) are largely standardised, based on the definition of Ecoregions within the region of focus. The greatest amount of variation in source data comes from the Biological data, where information on Hard coral and Macroalgae percent cover and the relative abundance of Fishes can involve a number of different methodologies, taxonomic codes and derived measures.
As part of GCRMN's mission to facilitate the integration of coral reef monitoring data from around the globe, the GCRMN has adopted developed a of data tiers to accommodate these different levels of taxonomic detail, sampling designs and spatial and temporal accuracy. These data standards are outlined in the GCRMN's Governance & Implementation Plan.
The tiers of this data quality model allow for data to be utilised for different purposes and underlying assumptions, while retaining sufficient detail for dedicated analyses.
This wiki page outlines the basic structure for Biological data collected as part of coral reef monitoring that reflect the GCRMN standard. Following these guidelines can greatly facilitate the standardisation and summary of raw_data for the RLE evaluation in this repository and provides some guidance on how to apply these standards for this project.
Key elements of the GCRMN data quality model
The GCRMN data quality model includes scores for team experience, sample design, level of evidence, and documentation, including:
- number of sampling units, of specified types
- physical grain of unit measure (ie. spacing/size issues)
- taxonomic identification level (how precise)
- experience and training of the monitoring team
For the application of the RLE framework for coral reefs, this project focuses on #1 "sampling units", #2 "physical grain" and #3 "taxonomic identification".
Variables and Data Standards
A key variable required for Goal 1 of the GCRMN is Hard coral cover and composition, which includes different levels of data quality:
- Level 1 "Live hard coral" as a single indicator
- Level 2 "Live hard coral" disaggregated by functional group (i.e. according to morphology and growth form)
- Level 3 "Live hard coral" disaggregated by genus or by species
As part of the characterisation of reef status and trends, information on Fleshy algae cover provides a valuable indicator of coral reef health. This variable includes the data quality levels of:
- Level 1 "Fleshy, macro and turf algae cover" as a single indicator (i.e. total % cover)
- Level 2 "Fleshy, macro and turf algae cover" disaggregated by primary functional groups based on morphology (e.g. turf, cruostose)
- Level 3 "Fleshy, macro and turf algae cover" disaggregated by genus or higher resolution of functional groups, including information on canopy height and/or biomass.
The CGRMN coral reef monitoring can include these categories:
Lastly, Fish abundance and diversity is an important variable that characterises the state of and processes that maintain reef health (e.g. grazing). There is a range of techniques and expertise required to monitor this variable, including a subset of key functional groups, inclusion of fish sizes, and abundance. An indicative guide of the data quality levels include:
- Level 1 "Basic" data collection of a subset of key families reported without sizes
- Level 2 "Intermediate" data with a comprehensive range of families and fish sizes
- Level 3 "Advanced" data where fish data are collected at the species level for the majority of families present, size class information, and replicate transects per site
General application for this project
The general strategy for managing the different levels of detail is to retain information from finer levels of detail as much as possible, and then aggregate to coarser levels when necessary. For this project, example analyses are based on the coarser level (i.e. Level 1) to maximise the data coverage spatially and temporally.
The code can be modified to summarise data collected at a finer resolution (i.e. Level 2 or Level 3).
Similarly, different levels of sampling (e.g. quadrates, transects) will be aggregated and summarised for a site average.