Fig 1.
Schematic diagram of the different steps undertaken within the analysis.
Multiple data sources were combined (+), compared relative to each other (/) and tested against defined criteria (?). Colors match respective Methods sections: (i) Recreation use data curation (green); (ii) Water status and geospatial data (blue); (iii) Null models (orange); (iv) Statistical analysis (purple).
Table 1.
Data obtained from individual River Basin Management Plans (Environment Agency, 2015) and Land Cover Map of Great Britain 2007 [31].
See inset in Fig 2 for location of River Basic Districts.
Fig 2.
Available datasets for cultural ecosystem services use in rivers, canals and lakes across England.
Geo-referenced visitation data from the Managing Engagement with the Natural Environment (MENE, Natural England 2009–2014; n = 4459); boating visits in the Watersports Participation Survey (British Marine, MCA, RNLI, RYA, British Canoeing and CEFAS 2014; n = 1298); fishing sites on fishinginfo.co.uk (Angling Trust 2015; n = 816); and outdoor swimming sites on wildswim.com (Outdoor Swimming Society 2015; n = 565). Inset shows the locations of the eight River Basin Districts in England (north to south): Northumbria (NB), North West (NW), Humber (HU), Anglian (AN), Severn (SV), Thames (TM), South East (SE) and South West (SW). Only points near (≤1km) of a river body with a reported ‘overall water status’ (i.e. WFD water quality standard) in 2014 were included.
Table 2.
Null models for the expected % of visits in good/high overall water status sites.
Table 3.
Number of visits and percent in good/high overall water status sites in all eight River Basin Districts.
Fig 3.
Association of good/high overall water status and use of cultural ecosystem services for the eight River Basin Districts.
The Odds Ratio (OR) of each River Basin District measures the likelihood that actual visits take place in sites characterized by good/high overall water status compared to random locations selected under a null model accounting for demand and substitutability (Table 2). OR exhibits a statistically significant positive (negative) association (i.e. visits in good/high overall water status sites are more (less) common than random; solid colours) if the 90% confidence interval is completely above (below) the line OR = 1. The robustness of the results is tested by comparing null models, including a null model without weighting. See Fig 2 for River Basin Districts acronyms.
Table 4.
Number (and codes) of River Basin Districts agreeing with the water quality—recreational ecosystem services hypothesis (or variants thereof): if good water quality is important for recreational ecosystem services we expect either (a) a positive association with water quality for all services or (b) positive association at least for services with significant direct contact with water—swimming and fishing, (c) stronger association with water quality for swimming and fishing relative to walking and boating.
The p-values denote the probability of getting equal or fewer RBDs meeting the criteria by chance alone, assuming a binomial distribution with 0.5 probability of success per trial.