The objective of this project was to improve understanding of the hydroclimatic drivers of biotic communities and ecosystem processes in ephemeral stream channels of the southwestern United States as a basis for projecting response to the changing regional climate. Guiding questions were: (1) How do catchment attributes, precipitation, and position in the stream network predict duration of stream flow? (2) How do community structure and function vary: (i) across a climatic aridity gradient; (ii) among streams that differ in stream flow duration and presence of shallow groundwater; (iii) among channel, riparian, and upland positions; and (iv) between wet and dry seasons?

A space-for-time substitution approach was followed in which changes across spatial water gradients are considered representative of anticipated changes in time under the climate change scenario of increased aridity. Sixteen stream sites were selected that spanned aridity zones (arid, semiarid, semihumid), stream flow permanence (ephemeral to perennial), and location within the stream network (piedmont, canyon, or alluvial basin). Electrical resistance sensors and USGS stream gauge data were used to quantify stream flow duration and stream water presence. Flow data and stream channel sediment data were used to estimate saturated hydraulic conductivity and potential annual infiltration. Groundwater monitoring wells were installed at streams in different landscape settings. Seasonal and annual variability in community composition and ecosystem processes were characterized via biannual sampling from 2010 to 2012. Data were collected on vegetation volume, above-ground herbaceous biomass, ground cover, and alpha species diversity. Soil seed banks were assessed with the seedling emergence method. Soil nutrient dynamics and release, among other standard soil physical and chemical characterizations, were assessed by measuring exchangeable nitrogen extracts and resins. Using grey oak (Quercus grisea) and Arizona sycamore (Platanus wrightii) litterbags, decomposition rates were assessed over an 18-month period. For arthropods, pitfall traps were set during dry and wet seasons in channel, riparian, and upland locations in canyon and piedmont reaches to measure spatio-temporal variation in alpha- and beta-diversity. Monthly collections of arthropods were used to measure secondary production.

Annual stream flow duration of the 16 stream sites ranged widely, attributable to variation in precipitation, temperature, and stream density in the catchment. Five flow regimes were defined based on flow and water duration: dry-ephemeral, wet-ephemeral, dry intermittent, wet intermittent, and semiperennial. Relationships between magnitude and temporal distribution of rainfall and stream flow permanenance varied by stream flow type. Duration of flow and precipitation were decoupled for intermittent and semiperennial streams because of contributions from groundwater discharge and the vadose zone. By contrast, stream flow at ephemeral sites responded rapidly to rainfall and overland flow. Infiltration rate (except where flow was perennial) was directly related to duration of stream flow.

Riparian vegetation was influenced by stream flow duration and water table presence, seasonal rain and flood pulses, and direct effects of aridity. As aridity increased among the ephemeral stream sites, the riparian zone had increasingly less vegetation volume, fewer plant species, and greater relative abundance of woody (versus herbaceous) vegetation. Further, the riparian vegetation and the matrix vegetation became increasingly dissimilar. Deep-stored flood water, and direct precipitation, maintained the combination of trees and herbaceous plants, respectively, at ephemeral streams. Within the semihumid Huachuca Mountains, seasonal fluvial disturbance from the strong monsoonal floods increased evenness of the herbaceous vegetation. Of note, introduced Lehmann’s lovegrass (Eragrostis lehmanniana) was dominant only in the uplands and not in the riparian zone of the Huachuca Mountain ephemeral streams.

Along the continuum from ephemeral to semiperennial stream flow, trade-offs were apparent between riparian forest biomass and plant species diversity. Woody plant biomass provides the main structure in riparian communities, and was linked with a shallow water table and high stream flow duration. Herbaceous plants provide most of the species diversity, and their response was decoupled from that of trees: herbaceous cover and richness decreased with stream flow duration (owing to light limitation) and, at sites with sparse tree cover, seasonally pulsed with monsoon rains and floods (and to a lesser extent with winter precipitation). Because herbaceous species were numerous, the net effect was a decline in total plant species richness at the wetter sites. The numbers of species in soil seed banks showed an interaction between site elevation and aridity. For ephemeral streams, seed banks had increasingly more species as aridity decreased (and precipitation increased), similar to patterns shown by terrestrial vegetation. Soil seed bank species numbers in riparian zones of intermittent streams showed a reverse pattern, given that few herbaceous species were present in the densely canopied streams of the high elevation semihumid sampling areas.

Decomposition and nutrient release were tightly coupled to stream flow and water presence only for the narrow band immediately associated with the stream channel. Leaf decomposition rates were higher in channels than in riparian zones, and for the channel position, were higher in wet-intermittent and semiperennial sites than in those with less frequent flow. Cumulative days of soil-water presence emerged as a significant explanatory variable for rate of decomposition. Surprisingly, rate of decomposition did not differ between the riparian zone and adjacent upland, perhaps owing to regulation by precipitation. Nitrate availability and release were higher in the riparian and upland positions compared to channel positions and, in the channel position only, were highest at ephemeral sites.

The production and community composition of ground dwelling arthropods were strongly influenced by flow permanence, with effects being seasonally dynamic. Median secondary production was positively related to annual stream flow presence and was consistently high in the wetter stream sites (i.e., most canyon sites). Peak secondary production, however, was highest in the ephemeral streams (including piedmont settings) but was highly transitory, pulsing during the one to two month period following monsoon rainfall. Alpha-diversity and beta-diversity (total turnover across the channel-upland transition) were related to annual water presence but varied by season (positive and asymptotic in dry season, negative in monsoon).

This project’s results show the need to conserve and protect a variety of stream flow types, in a variety of locations, to meet the multiple (and sometimes mutually exclusive) goals of maintaining high ecosystem productivity and high species richness. They also show that individually and collectively, the many small, unnamed ephemeral streams in the piedmont of the Huachuca Mountains and Barry Goldwater Range have high conservation value. Maintaining many small ephemeral washes across the landscape will help to maintain regional diversity and help to buffer upland taxa from periodic drought.

Understanding the ways, and rates, in which different water sources influence the structure and function of temporary streams will help managers interpret and anticipate ecosystem changes arising from regional climate shifts. Riparian zones are ecotones between terrestrial and aquatic zones, and are influenced by processes associated with each type of ecosystem. For ephemeral streams, structure and function is strongly regulated by terrestrial processes (e.g., precipitation) and many elements of the vegetation will shift rapidly in response to precipitation changes owing to their capacity to regenerate from soil seed banks. Structure and function of semiperennial streams, in contrast, are strongly regulated by aquatic processes (e.g., surface and subsurface water flows). Such sites will be slow to respond, given their dominance by long-lived riparian trees, but also will be buffered from rapid hydrologic change owing to the slower movement of riparian water sources through the hydroclimatic system.

One anticipated consequences of climate change is more intense storms, hence increased scour of vegetation. The soil seed bank results have implications for management actions that focus on restoration of plant communities following disruption from events such as scouring floods, given that donor soils can be an effective restoration tool for restoring small-seeded, herbaceous plant species to depauperate sites.

Finally, the results highlight the need to document, map, name, and protect the many ephemeral channels and associated riparian zones on the piedmont of the mountains in semiarid and semihumid areas of the Arizona Sky Islands (including the Huachuca Mountains). These stream ecosystems are narrow and easily overlooked on drainage maps and in the field, given their similarity in vegetation structure to adjacent uplands. Despite this apparent structural similarity, they contribute disproportionately to ecosystem processes including decomposition of organic matter and to ecosystem structure by supporting diverse communities of plants and arthropods.