Skip to main content

Springs ecosystem classification supplementary information

Cite this dataset

Stevens, Lawrence (2020). Springs ecosystem classification supplementary information [Dataset]. Dryad.


Springs ecosystems are globally abundant, geomorphologically diverse, and bio-culturally productive, but are highly imperiled by anthropogenic activities. More than a century of scientific discussion about the wide array of ecohydrological factors influencing springs has been informative, but has yielded little agreement on their classification. This lack of agreement has contributed to the global neglect and degradation of springs ecosystems by the public, scientific, and management communities. Here we review the historical literature on springs classification variables, concluding that site-specific source geomorphology remains the most diagnostic approach. We present a conceptual springs ecosystem model that clarifies the central role of geomorphology in springs ecosystem development, function, and typology. We present an illustrated dichotomous key to terrestrial (non-marine) springs ecosystem types and subtypes, and describe those types. We identify representative reference sites, although data limitations presently preclude selection of continentally or globally representative reference springs of each type. We tested the classification key using data from 244 randomly selected springs of 13 major types that were inventoried in North America. The dichotomous key correctly identified springs type in 87.5% of the cases, with discrepancies primarily due to differentiation of primary versus secondary typology, and insufficient inventory team training. Using that information, we identified sources of confusion and clarified the key. Among the types that required more detailed explanation were hypocrenes, springs, in which groundwater is expressed through phreatophytic vegetation. Overall, springs biodiversity and ecosystem complexity are due, in part, to the co-occurrence of multiple intra-springs microhabitats. We describe microhabitats that are commonly associated with different springs types, reporting at least 13 microhabitats, each which can support discrete biotic assemblages. Interdisciplinary agreement on basic classification is needed to enhance scientific understanding and stewardship of springs ecosystems, the loss and degradation of which constitute a global conservation crisis.

We provide data on springs ecosystem classification using non-sensitive springs data selected in an unbiased fashion from Springs Online ( to test the springs classificaiton dichotomous key and description (attached as separate pages to this dataset).   


The data provided for analysis in Table 3 were selected in a stratified, quasi-random method from the information available in Springs Online to test the dichotomous classificaiton key. Non-sensitive [not specifically excluded from use by the springs steward(s)], were randomly selected in approximate proportion to springs type frequency within the Springs Online database. Springs data were used to test whether the springs classification key and description (also attached) correctly identified the field-determined springs type. Comparisons were performed by the authors and by Museum of Northern Arizona's Springs Stewardship Institute staff. The proportion of correct answers are presented in Table 3. Following this analysis, the dichotomous key was refined to improve its ability to identify springs geomorphic type.       

Usage notes

The site numbers listed refer to springs with substantial inventory data in Springs Online ( Access to the inventory data for each springs is unrestricted for many springs,but some require permission from the intellectual property owner (ususally the springs steward), which may be a governmental agency, a non-governmental agency, or a private individual. Please contact the Museum of Northern Arizona Springs Stewardship Institute (3101 N. Ft. Valley Rd., Flagstaff, AZ 86001 USA; (928) 774-5211 X 231; or for questions about, or assistance with access to these data.


No organizational funding was provided for this analysis., Award: None

No organizational funding was provided for this analysis., Award: None