Skip to contents

Validate the raw QWA data

Source: R/clean_raw_data.R
validate_QWA_data.Rd

Initial checks to ensure the quality of the raw QWA data. This function checks for the following issues:

  • undated images, i.e. YEAR is NA or in the future (raises error)

  • images without cell wall thickness estimates (raises warning, since CWT is required for conifers but may not be available in the case of angiosperms)

Usage

validate_QWA_data(QWA_data, df_meta, verbose_flags = FALSE)

Arguments

QWA_data

a list containing the cells and rings dataframes

df_meta

a dataframe containing the metadata for the images (spatial_resolution required for the incomplete innermost ring check)

Value

validated QWA_data (no changes to cells df, but added minimum required ring flag columns to rings df).

Details

Next, the function identifies the rings with the following issues:

  • incomplete rings

  • missing rings

  • duplicate rings

Here, by incomplete rings we mean those at the inner (pith) and outer (bark) boundaries of an images, which are cut-off by the image or slide border. For these rings, some cells are usually recognized but the MRW can NOT (outer) or NOT ACCURATELY (inner) be estimated. In some cases, the user may have manually deleted the incomplete rings within ROXAS already, so it is not a priori clear that all inner- and outermost rings per image are incomplete.

NOTE: Because ROXAS uses the outer ring boundary to estimate MRW, the innermost ring generally has an MRW estimate (that is not based on the true ring boundary but rather the image border), while the outermost ring has no MRW value (except if it is actually complete either because it is at the at the bark or because the user removed the incomplete ring manually in ROXAS). Therefore, we perform an additional check on the border shape and position to check if an innermost ring is incomplete, while the outermost ring is flagged as incomplete if and only if it has no MRW.

Missing rings are for years that have no discernible ring in the image, but have been manually added in ROXAS during cross-dating, leading to an entry in the rings data but no corresponding no entries (cells) in the cells data. This is usually the case with wedging rings.

Duplicate rings are those that are present in multiple images due to them overlapping. All years which have cells in more than one image are flagged and ranked by their number of cells. The (complete) year with the highest number of cells for each overlap is the one that would then usually be selected for further analysis when building chronologies.