Vernal pool, swale, and headwater stream complexes are fully integrated into broader flow networks during pronounced wet seasons (Fig. 6). The duration of hydrologic connection within these vernal pool, swale, and headwater stream complexes and to downstream waters varies widely as a function of precipitation, but is ~ 85 days in a normal year in the study area (Fig. 5). While connected, these vernal pool landscapes serve as a prominent source of streamflow to downgradient waters, as storm peaks commonly initiate first in the vernal pools and draining swales, and thereafter translate to headwater streams and sequentially downstream to the Sacramento River (e.g., Fig. 7).
Vernal pools, swales, and headwater streams connect to each other and to downstream waters as a strong function of precipitation, as a result of minimal surface water and shallow-groundwater storage (Figs. 3 and 4). Vernal pool landscapes are characterized by a lack of water storage capacity due to shallow microtopographic relief and the presence of slowly permeable horizons (Smith and Verrill, 1998; Rains et al., 2006, 2008). Local microtopographic relief and the cross-sectional areas of the wetlands and swales are both small, so surface water storage is limited (Fig. 3). Similarly, there typically are multiple slowly permeable horizons within the upper 1–2 m of the soil profile (Fig. 4). This limited surface water and groundwater storage capacity fills rapidly with the onset of the wet season and/or an individual storm (e.g., Rains et al., 2006). Surface water accumulates in the microtopographic lows, which on these vernal pool landscapes are the vernal pools, swales, and headwater streams, and surface water storage begins to reach capacity.
Once storage capacity is exceeded, subsequent precipitation initiates immediate runoff and inundation in the vernal pool, swale, and headwater stream complexes, which is then translated downgradient (Fig. 6). The vernal pool landscapes are nearly level to gently sloping, approaching 0% in some locations (Fig. 3). Additionally, the watershed areas of individual vernal pools are small. Consequently, surface water flows accumulating within these small watersheds lack sufficient stream power to either erode clear bed and bank features or develop dendritic flow networks. Instead, vernal pools and swales are first interconnected in deranged flow networks, with surface water flowing slowly among the vernal pools and swales, trending downgradient towards the high terrace scarps. Some vernal pools and swales only receive water from their immediate watershed area (i.e., “feeder” vernal pools, sensu Bauder [2005]) and other vernal pools and swales receive water from both their immediate watershed area and other upgradient vernal pools and swales (i.e., “collector” vernal pools, sensu Bauder [2005]) (Fig. 9). These waters then flow into headwater streams. As a result of these processes, each individual vernal pool and swale contributes differentially to flow generation, with contributions varying in both space and time.
As these waters flow off the high terraces and down the high terrace scarps, combined watershed areas are larger and slopes are steeper, resulting in flows with sufficient stream power to erode stream beds and banks and develop dendritic flow networks (e.g., Fig. 14 in Horton, 1945). Once these vernal pools and swales are connected to these headwater streams, they all become part of integrated flow networks. Over the course of the wet season and/or individual storms, these integrated flow networks contribute flow downstream, thereby comprising the variable source area of downstream waters (e.g., Dunne and Black, 1970; Hewlett and Nutter, 1970).
Vernal pool, swale, and headwater stream complexes exhibit a range of connectivity behavior over the course of the year as a result of the strongly seasonal Mediterranean climate (Fig. 10). Vernal pool, swale, and headwater stream complexes are not inundated during the dry season. They are seasonally and intermittently inundated, but not fully connected, as the wet season begins. As the wet season progresses and the limited soil water storage capacities are exceeded, they become inundated, fully connected, and flowing. As the wet season ends, and evapotranspiration processes are amplified with increasing seasonal heat, they become intermittently inundated and flowing.
Annual precipitation is highly variable in Mediterranean climates (e.g., Bonada and Resh, 2013). We overcame this by collecting data over a period of five years (i.e., October 2012-September 2017). During this time, precipitation ranged from 59%-114% of normal (Table 4; Fig. 5). This variability is typical, as precipitation ranged from 56%-187% of normal between 1981–2010 (Red Bluff Municipal Airport, Station 24216, 1981–2010).
The hydrographs of the vernal pool, swale, and headwater stream complexes responded rapidly and linearly to this precipitation (Figs. 6 and 7). Precipitation events, total precipitation volume, and subsequent flow in the vernal pool, swale, and headwater stream complexes tend to occur during concentrated time periods during the wet seasons in this physiographic region (Fig. 6). Therefore, the rapid responses followed by brief recessional limbs in the hydrographs typically link, producing ~ 85 days of continuous or near-continuous flow from the vernal pool, swale, and headwater stream complexes during a normal year (Fig. 6). As a result of the highly variable amount of annual precipitation in this region, flow from the vernal pool, swale, and headwater stream complexes also varies greatly from year to year. Flow from the vernal pool, swale, and headwater stream complexes ranged from 19–110 days during the time period analyzed in this study (Table 4; Fig. 5). If this relationship holds throughout the normal range of precipitation, then flow from the vernal pool, swale, and headwater stream complexes likely ranged from ~ 20–200 days per year between 1981–2010.
Though they constitute a small proportion of watershed area, vernal pool, swale, and headwater stream complexes play a disproportionate role in routing precipitation from the vernal pool landscapes to downstream waters during pronounced wet seasons. Vernal pool landscapes comprise 20%, 8%, and 13% of the Coyote Creek watershed, the Oat Creek watershed, and the total combined watershed areas, respectively (Fig. 2). Similarly, vernal pool, swale, and headwater stream complexes on these vernal pool landscapes comprise 4.0%, 1.6%, and 2.6% of the Coyote Creek watershed, the Oat Creek watershed, and combined watersheds, respectively (Fig. 2). Our estimates are consistent with other estimates throughout the Central Valley, including those that include the Coyote Creek and Oat Creek watersheds (e.g., Holland 1996; 1998).
Though vernal pool, swale, and headwater stream features are structurally distinct from one another, they are functionally similar and collectively serve as key sources of streamflow when fully connected and flowing during the wet season (Fig. 5). At such times, they act as part of the variable source area for the Coyote Creek and Oat Creek watersheds, as flow peaks appear first on the vernal pool landscapes and are thereafter propagated sequentially and continuously downstream to the Sacramento River (Figs. 6–8).
The convolution of flows contributed from the vernal pool, swale, and headwater stream complexes over space and time therefore contributes to the physical integrity of downstream waters. This logic is in line with the conceptual model proposed by Cohen et al. (2016) and extended specifically to flow generation by Rains et al. (2016), in which the convolution of these spatially and temporally varying contributions results in the characteristic natural flow regime, over a range of time periods (e.g., throughout an individual storm, during a single year, and over the course of many years). The cumulative contribution of the vernal pool, swale, and headwater stream complexes to flow generated within a watershed and the resulting hydrograph can be understood by the convolution of all flow and across all travel times to a downstream location (Fig. 9, 10). The space- and time -integrated contributions to flow generation contribute to the natural flow regime of that system, and in turn, contribute to the maintenance of the chemical, physical, and biological integrity of downstream waters. However, the magnitude of this contribution and the degree to which the size and spatial arrangement of the individual vernal pool, swale, and headwater stream complexes affects these relationships remains to be established.
The results of this study have important implications for land-use management and regulation of activities within vernal pool landscapes. For example, the lack of surface water and shallow-groundwater storage capacity is in part a function of the slowly permeable horizons in upper pats of modal soil profiles (Fig. 4), which is itself a function of the relatively old age of the deposits on which these vernal pool landscapes occur (Helley and Harwood, 1985; Smith and Verrill, 1998). These features are widely retained because regional land uses have long been little more than grazing of domestic livestock. Ecosystems of the Central Valley evolved in the presence of light to moderate grazing by native ungulates, like the Tule elk (Wagner, 1989). Many of the early European settlers in the Central Valley were beneficiaries of the Spanish Land Grants, in which settlers were granted large tracts of land on which they subsequently grazed cattle (Allen, 1935). Today, cattle grazing remains a widespread and economically important land use in the region (Tehama County Department of Agriculture, 2021). However, irrigated agriculture is becoming increasingly dominant, and the two highest value agricultural commodities for the region are now walnuts and almonds, respectively (Tehama County Department of Agriculture, 2021). These deep-rooted trees cannot be grown on these soils with slowly permeable horizons in the upper parts of modal soil profiles. Therefore, to allow production of these crops, soils are first prepared by deep ripping, in which deep working tines are pulled behind heavy equipment to mechanically break up and shatter the slowly permeable horizons, typically including the duripan. This tillage treatment then increases the depth of both water infiltration and root penetration into the subsurface (Hussein et al., 2019). Though data are sparse, deep ripping is specifically designed to have profound and immediate effects on the patterns and rates of the downward movement of water and water storage in ripped soil profiles. Therefore, the result of deep ripping is to have significant and discernable impacts on the local scale patterns of water flow and circulation. At landscape scales, deep ripping significantly impacts processes of flow generation and the resulting downstream natural flow regimes.
The results of this study also have important implications for policy, especially as it relates to the definitions of the geographic extent of Waters of the United States (WOTUS) that are subject to regulation under the Clean Water Act (CWA). The overall goal of the CWA is to “restore and maintain the chemical, physical, and biological integrity of the Nation’s waters” (33 U.S.C. § 1251 et seq. [1972])). Crucial to achieving this goal is the role of various definitions of WOTUS as they relate to the establishment of U.S. federal jurisdiction. The federal government has long struggled with the definition of WOTUS. In recent years alone, the Obama administration promulgated the 2015 Clean Water Rule (CWR) (Federal Register Vol. 80, No. 124) which was subsequently rescinded by the Trump administration in 2019 (Federal Register Vol. 84, N0. 204), and the Trump administration promulgated the 2020 Navigable Waters Protection Rule (NWPR) (Federal Register Vol. 85, No. 77) which was subsequently vacated by the courts in Pasqua Yaqui Tribe v. U.S. Environmental Protection Agency (2021). Throughout the debate, tributaries to “traditionally navigable waters” (TNWs) have consistently been included as WOTUS. But where do those tributaries begin? Our results show that vernal pool, swale, and headwater stream complexes are important parts of tributary networks, being connected to downstream waters both structurally (e.g., the elements are physically connected to one another, and to downstream waters) and functionally (e.g., water flows between the elements, and to downstream waters (e.g., Sullivan et al., 2019, 2020)). These vernal pool, swale, and headwater stream complexes perform lag (i.e., detention), sink (i.e., retention), and source (i.e., transmission) functions (Rains et al., 2016), and therefore contribute to the development and maintenance of natural flow regimes (Poff et al., 1997). These contributions, in turn, help “restore and maintain the chemical, physical, and biological integrity of the Nation’s waters.” In this case, those “Nation’s waters” include lower Coyote Creek and Oat Creek, both designated critical habitat for the U.S. federally listed (threatened) Central Valley spring-run Chinook salmon (Oncorhynchus tshawytscha) and U.S. federally listed (threatened) Central Valley steelhead (Oncorhynchus kisutch).