PERSPECTIVES The Stream Biome Gradient Concept: factors controlling lotic systems across broad biogeographic scales Walter K. Dodds1,5, Keith Gido1,6, Matt R. Whiles2,7, Melinda D. Daniels3,8, and Bartosz P. Grudzinski4,9 1Department of Biology, Kansas State University, Manhattan, Kansas 66506 USA 2Department of Zoology, Southern Illinois University, Carbondale, Illinois 62901 USA 3Stroud Water Research Center, 970 Spencer Road, Avondale, Pennsylvania 19311 USA 4Department of Geography, Kansas State University, Manhattan, Kansas 66506 USA Abstract: We propose the Stream Biome Gradient Concept as a way to predict macroscale biological patterns in streams. This concept is based on the hypothesis that many abiotic and biotic features of streams change predict- ably along climate (temperature and precipitation) gradients because of direct influences of climate on hydrology, geomorphology, and interactions mediated by terrestrial vegetation. The Stream Biome Gradient Concept gener- ates testable hypotheses related to continental variation among streams worldwide and allows aquatic scientists to understand how results from one biome might apply to a less-studied biome. Some predicted factors change monotonically across the biome/climate gradients, whereas others have maxima or minima in the central portion of the gradient. For example, predictions across the gradient from drier deserts through grasslands to wetter forests include more permanent flow, less bare ground, lower erosion and sediment transport rates, decreased importance of autochthonous C inputs to food webs, and greater stream animal species richness. In contrast, effects of large ungulate grazers on streams are expected to be greater in grasslands than in forests or deserts, and fire is expected to have weaker effects in grassland streams than in desert and forest streams along biome gradients with changing precipitation and constant latitude or elevation. Understanding historic patterns among biomes can help describe the evolutionary template at relevant biogeographic scales, can be used to broaden other conceptual models of stream ecology, and could lead to better management and conservation across the broadest scales. Key words: stream, biome, lotic, macro-scale, macrosystems, biogeography Stream ecologists have devised several ecological concepts criteria and nutrient ecoregions (Omernik 1987) and geo- to understand streams at broad spatial or temporal scales. graphic patterns of animal communities across freshwater Transitional characteristics of stream ecosystems across a ecoregions (Abell et al. 2008). Species distributions and gradient from headwaters to large rivers were linked in the state and national jurisdictions occur across very large River Continuum Concept (RCC; Vannote et al. 1980). scales that often include multiple biomes. Some attention This concept centered on forested streams, but the au- has been paid to how river networks might vary across thors also considered that different biomes could fitinto broad scales (e.g., McCluney et al. 2014), but we are not the concept with some modification. Other synthetic ap- aware of a specific framework that considers how streams proaches to stream and river ecology, e.g., the Flood–Pulse vary predictably across the broadest scales among differ- concept (Junk et al. 1989) and the Riverine Ecosystem ent biomes. Our research on multiple stream biomes has Synthesis (Thorp et al. 2006), consider terrestrial influ- led us to view gradients across multiple biomes as hav- ences, but not with primary emphasis on the biomes in ing the potential to predict differences in general stream which the rivers or streams are embedded. community and ecosystem characteristics at the broad- Very large-scale approaches (e.g., whole large river ba- est scales. sins, continents, cross-biome) are important for many key Control of plant communities by climatic (temperature ecological issues (Heffernan et al. 2014), including nutrient and precipitation) gradients has been recognized for >65 y E-mail addresses: [email protected]; [email protected]; [email protected]; [email protected]; [email protected] DOI: 10.1086/679756. Received 22 April 2014; Accepted 30 October 2014; Published online 20 January 2015. Freshwater Science. 2015. 34(1):1–19. © 2015 by The Society for Freshwater Science. 1 2 | Stream Biome Gradient Concept W. K. Dodds et al. and forms the basis of the biome concept (Holdridge transitions (e.g., position of major atmospheric convec- 1947). We extended the biome concept to investigate how tion cells), are common in many parts of the world and gradients of precipitation and temperature can predictably lead to gradients from deserts to forests. These patterns control differences in stream hydrology and geomorphol- can span continents or occur over relatively short distances ogy directly or as a consequence of interactions of streams in mountainous areas. Coupled with these patterns, tem- with the terrestrial biomes that dominate under specific perature gradients associated with elevation or latitude in- climatic conditions. We further explored how interbiome fluence vegetative characteristics and hydrology. These differences can constrain the structure and function of climate patterns can lead to gradients from warm tropical stream communities and ecosystems. lowlands to arctic tundra, or over shorter distances from Holdridge (1947) showed how interactions between tem- tropical lowland to tundra on high tropical mountains. perature (as influenced by latitude or elevation) and pre- We asked whether stream properties change continuously cipitation could be used to predict the functional groups of across large-scale biome gradients, are invariant, or have vegetation found in terrestrial environments, and this ap- unique attributes dictated by the biomes in which they oc- proach was adopted by others, e.g., in the biome continuum cur that do not scale linearly with climate gradients (e.g., concept of McIntosh (1967). We extended this view to pro- nonmonotonic patterns across gradients). Freshwater bi- pose the Stream Biome Gradient Concept. This concept is omes (ecoregions) have been assigned unique status as based on the hypothesis that streams change predictably ecoregions based on vertebrate inhabitants (Abell et al. along climate (temperature and precipitation) gradients be- 2008), but the relationships of these ecoregion designations cause of direct influences of climate on hydrology and geo- to other stream properties and to terrestrial biomes at con- morphology and indirect influences mediated by terrestrial tinental scales are not clear. vegetation. We use this concept to predict stream abiotic Streams are the dominant interface between terrestrial and biotic characteristics (e.g., hydrology, geomorphology, and aquatic ecosystems (Hynes 1975). They transport and water quality, ecosystem metabolism, and animal diversity transform (Mulholland et al. 2008) materials from land to and function) across broad climate gradients (Fig. 1). We the ocean and ultimately determine coastal marine pro- developed this idea as we tried to understand and describe ductivity. Streams and rivers also are important sources of grassland streams in relation to other stream types on greenhouse gasses (Beaulieu et al. 2011, Raymond et al. which we have worked (e.g., deserts, tropical rain forest, 2013), but we do not know how these functions are dis- deciduous, and evergreen temperate forests). tributed globally and relate to biomes. Precipitation gradients across continental land masses, The direct link between terrestrial and lotic ecosystems which often are related to rain shadows or other climatic is runoff. Temperate forests produce 30% of global runoff. Figure 1. Climate zones, vegetation types, and relationship to stream characteristics (modified from Holdridge 1947). Volume 34 March 2015 | 3 In contrast, tropical and subtropical evergreen forests pro- tem properties, including water quality, nutrient dynamics, duce ∼50%, and grasslands and savannas the remaining and metabolism, as they vary with biome, 4) variation in 20%, of runoff. Deserts contribute very little runoff,though animal community diversity and dynamics, and 5) linkages they constitute a substantial area (Table 1). Streams in with other integrative ideas on lotic ecology. We caution temperate forests appear to be the most studied. We ana- that we are addressing very broad biogeographic patterns. lyzed ISI Web of Science (Thomson Reuters, Philadel- These generalizations are not meant to apply to smaller phia, Pennsylvania) citation® records (January 2013) by scales (e.g., strong rain shadows and elevation gradients on searching for papers in the area of “ecology” with the islands), although some of the concepts given here may terms “river” or “stream” in them. Within these results, well transfer to some aspects of smaller-scale gradients. most citations included the term “forest” (8609) followed Many of the topics discussed herein are not yet well stud- by “tropic*” (594; * is a wild-card search term), “rain forest” ied in all systems, so we were able to make predictions (156), “arctic” (153), “grassland” (166), “prairie” (77), or and hypothesize based only on limited published informa- “tundra” (46). This approach is a coarse way to evaluate tion or broad, well established biological and physical the literature, but it does support the idea that knowledge principles. of stream ecology is strongly influenced by research from forested streams. Streams types are not studied proportion-