What's in a Name? Patterns, Trends, and Suggestions for Defining Non-Perennial Rivers and Streams

Abstract

Rivers that cease to flow are globally prevalent. Although many epithets have been used for these rivers, a consensus on terminology has not yet been reached. Doing so would facilitate a marked increase in interdisciplinary interest as well as critical need for clear regulations. Here we reviewed literature from Web of Science database searches of 12 epithets to learn (Objective 1-O1) if epithet topics are consistent across Web of Science categories using latent Dirichlet allocation topic modeling. We also analyzed publication rates and topics over time to (O2) assess changes in epithet use. We compiled literature definitions to (O3) identify how epithets have been delineated and, lastly, suggest universal terms and definitions. We found a lack of consensus in epithet use between and among various fields. We also found that epithet usage has changed over time, as research focus has shifted from description to modeling. We conclude that multiple epithets are redundant. We offer specific definitions for three epithets (non-perennial, intermittent, and ephemeral) to guide consensus on epithet use. Limiting the number of epithets used in non-perennial river research can facilitate more effective communication among research fields and provide clear guidelines for writing regulatory documents.

Keywords: ephemeral; intermittent; latent Dirichlet allocation; literature review; non-perennial; river; stream; synthesis; temporary; text mining.

Figure 1.

Proportion of epithets by Web of Science (WoS) category. Displayed are the 40 categories (out of 148 unique WoS categories) that contained at least 50 papers across all 12 epithet corpora on the y-axis. The proportion of each category made up of epithets is on the x-axis. Each color represents a different epithet, with the most dominate epithet per category highlighted in black. The total number of papers appearing in each category is displayed as the sample size along each column for reference (environmental sciences had the largest number with 2398 while multidisciplinary chemistry had the smallest with 55 papers).

Figure 2.

A non-metric multidimensional scaling (nMDS) ordination (stress = 0.1029) representing similarities among the six topics found from latent Dirichlet allocation (LDA) topic modeling (blue words) and how they relate to papers about each epithet (black words). The large sample size made plotting each paper too cluttered, so average probabilities of the topic appearing in papers across all twelve corpora were used to understand broad, topical terms between epithets.

Figure 3.

Number of total papers published (A) and proportion of papers published under different epithets (B) over time frames. Note the large jump in publication rates after 1990, and the steady increase until 2016. Papers from all 12 epithets are published between 1996 and 2000, though “non-perennial” does not appear in 2001–2005, all 12 are again represented in results from 2006 onward.

Figure 4.

Publications of each epithet over time. Abstracts from 1990 to 2017 were compiled to see how epithet use has changed over time. Each epithet was counted once per abstract. Broader terms such as “seasonal,” “dry,” and “temporary” have been used the most frequently. “Temporary,” “non-perennial,” and “intermittent” are the three terms with the largest rate of increase over time.

Figure 5.

An nMDS representing the Euclidean distance among epithets and definition themes (stress = 0.0962). Epithets in blue text, definition themes in larger, black text. The section of the nMDS with overlap includes epithets “intermittent”, “temporary”, and “non-perennial” as well as the theme Phases of Drying: No Flow. Full theme list: predictability = predictability/seasonality; variability = variability/unpredictability; time frame = specific time frame mentioned; landform/climate = related to specific landform/climate; extremes = related to extremes (droughts and floods); pod = phases of drying: (Table S2).