Function to develop a pooling group based on catchment descriptors
Pool(CDs, N = 800, UrbMax = 0.03, DeUrb = TRUE, exclude = NULL, include = NULL)A data.frame of the pooling group with site reference row names and 24 columns, each providing catchment & gauge details for the sites in the pooling group.
catchment descriptors derived from either GetCDs or CDsXML
minimum Number of total gauged record years for the pooling group
Maximum URBEXT2015 level with a default of 0.03. Any catchment with URBEXT2015 above this level will be excluded from the pooling group
logical argument with a default of TRUE. If TRUE, the LCVs of all sites in the pooling group are "De-Urbanised".
sites to exclude from the pooling group. Either a single site reference or a vector of site references (numeric). If this is used the next site with the lowest SDM is included such that the total sample of AMAX is at least N.
sites to include that otherwise would not be included by default. For example if it is a subject site that has URBEXT2015 above UrbMax. Or one that has not been selected automatically using the similarity distance measure.
Anthony Hammond
A pooling group is created from a CDs object, derived from GetCDs or CDsXML, or specifically with the catchment descriptors (see arguments). To change the default pooling group, one or more sites can be excluded using the 'exclude' option, which requires either a site reference or multiple site references in a vector. If this is done, the site with the next lowest similarity distance measure is added to the group (until the total number of years is at least N). Similarly a site can be included specifically by using the include argument. Sites with URBEXT2015 (urban extent) > 0.03 are excluded from the pooling group by default. This threshold can be adjusted with UrbMax. If DeUrb is set as TRUE (the default), the LCV values for sites in the pooling group are de-urbanised. If the user has more data available for a particular site within the pooling group, the LCV and LSKEW for the site can be updated after the group has been finalised using the LRatioChange function.
The pooling method is as specified by FEH2025. The de-urbanisation functionality assumes that the growth curve associated with an annual maximum flow sample is impacted by urbanisation and that this impact can be modelled as a function of the catchment URBEXT. The method for pooling the catchments together is based on the similarity of AREA, SAAR, FARL, FPEXT, and BFIHOST. These were seen to have the most significant impact on the LCV and LSKEW - and ultimately to provide the lowest 'Pooling Uncertainty Measure' (a statistic for assessing the similarity between pooled and single site gauged estimates).
# Get some catchment descriptors
cds_73005 <- GetCDs(73005)
# Set up a pooling group object called pool_73005 excluding sites 79005 & 46003
# Then print the group to the console
pool_73005 <- Pool(cds_73005, exclude = c(79005, 46003))
pool_73005
Run the code above in your browser using DataLab