A spatially explicit framework for assessing the effects of weather and water rights on streamflow

Applied Geography 67 (2016) 14e26

Contents lists available at ScienceDirect

Applied Geography journal homepage: www.elsevier.com/locate/apgeog

A spatially explicit framework for assessing the effects of weather and water rights on streamflow Matthew J. Deitch a, *, Mia van Docto a, Shane T. Feirer b a b

Center for Ecosystem Management and Restoration, Oakland, CA, USA University of California Cooperative Extension, Hopland Research and Extension Center, Hopland, CA, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 2 October 2015 Received in revised form 25 November 2015 Accepted 27 November 2015 Available online xxx

Policies that establish availability of water for human uses are key tools for ecologically sustainable water management. In places where water management is decentralized and dispersed across a catchment, spatially explicit tools are essential for evaluating the cumulative effects of many instream diversions on discharge and water availability through the drainage network. We developed a spatially explicit model to evaluate the cumulative impacts of surface water rights under a decentralized management regime, and apply this framework to the Navarro River catchment, in northern California, and evaluate impacts relative to formal policies to determine whether water is available for further appropriation given environmental needs. Model results show that upstream water rights comprise a small fraction of normal-type winter discharge; but they comprise a much larger portion in dry years and all of discharge during the summer dry season. In a normal-type rainy winter season, water rights comprise less than 5 percent of the average discharge during the through almost all of the drainage network: by policy standards, water is widely available for further appropriation during this winter period. Most stream reaches where upstream water rights comprise more than 10 percent of discharge are small headwater streams (zero- and first-order). During other periods such as November and April (the beginning and end of the rainy season, respectively), impairment by water rights remains low under average conditions; but under dry conditions, most streams order 0e2 with upstream water rights are impaired more than 10 percent (with some more than 50 percent). During average summer conditions, 45 percent of the drainage network with upstream water rights is impaired by more than 20 percent; and 25 percent of these streams are impaired by more than 50 percent. These results indicate that water is available for further appropriation in winter; but water rights can cumulatively impair discharge by more than ten percent in small and large streams under dry-type conditions, and by more than 50 percent basinwide during the dry season. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Spatially explicit watershed model Cumulative effects Water availability Sustainable water management Environmental flows Mediterranean climate variability

1. Introduction Managing water to meet human and environmental needs is an increasingly common water resources paradigm in the 21st Century. The benefits of protecting streamflow and groundwater for environmental purposes are well-documented: they maintain habitat and processes necessary for sustaining aquatic and riparian organisms, and they sustain ecosystem services that benefit human well-being through the flow of goods and services from nature to people (Costanza et al. 1997; De Groot, Wilson, & Boumans, 2002;

* Corresponding author. E-mail address: [email protected] (M.J. Deitch). http://dx.doi.org/10.1016/j.apgeog.2015.11.018 0143-6228/© 2015 Elsevier Ltd. All rights reserved.

Wu, 2013). Requirements to maintain streamflow dynamics such as environmental flows (described by Arthington and Pusey (2003) as those streamflow dynamics that maintain geomorphic and ecological processes) are increasingly seen as necessary to protect the processes that sustain aquatic biota and public trust resources (Poff et al. 2010; Richter, 2010). In places where water is regulated by large reservoirs, managing streamflow to protect environmental needs is relatively straightforward. The flows necessary to sustain environmental needs can be determined through modeled outcomes or measured empirically (Tharme 2003), and once the dynamics of reaching and sustaining those flows have been agreed upon, streamflow can be released from dams according to prescribed flow release schedules (Postel & Richter, 2003; Richter & Thomas, 2007; Richter, Warner,

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

Meyer, & Lutz, 2006). Models have been developed to identify the dynamics of flows necessary to provide the natural variability that would occur in the absence of a large reservoir capable of regulating all streamflow dynamics (Poff et al. 2010; Richter, Baumgartner, Powell, & Braun, 1996, 2012; Richter, Davis, Apse, & Konrad, 2012). If water is not managed from a central location, protecting flows for environmental needs is more complex. Agricultural producers, rural residents, recreational operations, and other types of land managers outside of the areas served by large water providers often obtain water from streams, springs, and groundwater through individual methods (Antonino et al. 2005; Deitch, Kondolf, & Merenlender, 2009a; Habets, Philippe, Martin, David, & Leseur, 2014; Liebe, van de Giesen, & Andreini, 2005; Malveira, de Arujo, & Guntner, 2011; The Economist, 2007). The resulting water management regime is decentralized, managed by individual water users rather than a central water supplier; it results in small dispersed impacts spatially distributed through a drainage network, rather than focused on one particular location (Deitch, Merenlender, & Feirer, 2013). Surface water diversions under a decentralized management regime are smaller than those that occur at large water supply reservoirs because they only provide water to meet the needs of one or a few water users, but depending on where and when they occur, such small diversions can have substantial effects on streamflow locally where they operate (Deitch, Kondolf, & Merenlender, 2009b). The majority of the area in northern Coastal California is under a decentralized water management regime. Outside of a few large municipalities and small agricultural areas with water providers, people who live and work along streams obtain water through individual methods. Because of low abundance and poor quality of groundwater, and seasonal precipitation trends that result in virtually no rainfall during summer months, water is commonly diverted from streams and adjacent shallow aquifers to meet water needs through the dry season (Deitch & Kondolf, 2015). This means of obtaining water results in a direct ecological conflict: juvenile steelhead and coho salmon (Oncorhynchus mykiss and O. kisutch, respectively) also rely on water through the summer months to rear before migrating to the ocean. The common need for water through summer can result in adverse impacts to the streamflow that provides oversummering habitat for these juvenile fishes in the summer dry season (Deitch et al. 2009b; Grantham, Newburn, McCarthy, & Merenlender, 2012). Changes in environmental laws in the 1990s (in particular, the listing of steelhead and coho salmon under the federal Endangered Species Act and a state court-ordered directive to protect public trust resources) prompted the California State Water Resources Control Board (State Water Board) to develop a Policy for Maintaining Instream Flows in Northern California Coastal Streams (Policy) to address this seasonal conflict between human and ecosystem water needs (SWRCB, 2010). To provide adequate ecosystem protections, the Policy prohibits granting water appropriations during summer, instead only allowing new water appropriations over a defined winter period. Land managers seeking a new water appropriation thus need to store water onsite in tanks or small reservoirs in winter for use in summer. Additional restrictions were placed on new water appropriations to ensure adequate protections for winter flows (discussed in more detail below), but overall, the Policy assumes that new instream diversions in winter can be accommodated while still offering adequate ecological protections that diversions during other times of year cannot. The application of this new Policy in coastal California has led to many concerns and questions about the limitations it imposes. Water users and agricultural interest groups have voiced concern that the concentration of diversion to winter-only may, in some

15

cases, affect senior water right holders; and in others, prevent new water right applicants from obtaining water they need for agricultural production. Environmental advocates have expressed concern that the cumulative effects of many small diversions could impair flow in winter beyond what could be predicted and cause widespread impacts throughout the drainage network. Because of the decentralized nature of water management in this region, these questions are not easily answered: the effects of diversions on streamflow will vary through the drainage network, as well as through the year from rainy winter to dry summer. Spatially explicit GIS tools that can integrate cumulative effects of spatially distributed impacts and variations in discharge over space and time are critical for answering these important questions. In the research that follows, we use a spatially explicit basinscale model to calculate the cumulative impacts of existing water rights on discharge through the Navarro River drainage network, as well as, according to the Policy, whether and where additional water is available for further appropriation. Because of the importance of climate variability in understanding humaneecosystem interactions, we also evaluate effects of water rights on discharge during other times of year and under dry-year types to compare impacts under different climatic conditions. 2. Methods 2.1. Study area The Navarro River catchment in Mendocino County, California, drains a mostly forested 800 square km catchment (with some chaparral and grassland used for pasture) and reaches the Pacific Ocean 190 km northwest of the San Francisco Bay (Fig. 1). Over the past century, the predominant uses of land have been for timber harvesting in the conifer forest and pasture in the chaparral and grassland; in recent decades, some of the pasture has been converted to wine grape vineyards and apple orchards (covering approximately 1010 ha). The climate of the Navarro is typical of coastal California: it is characterized as Mediterranean, with most precipitation occurring as rainfall during wet winters and very little occurring during the dry summer. PRISM precipitation models estimate that rainfall in the Navarro catchment ranges from 1000 mm to 1500 mm in an average year; approximately 90 percent of the rainfall occurs during the wet half of the year (November through April). Less than 3 percent occurs during July, August, and September. Streamflow in the Navarro River and its major tributaries mirrors this pattern, alternating between an abundance of flow and frequent flooding through winter to approaching or reaching intermittence in summer (Fig. 2). The long summer dry season necessitates irrigation for agricultural production (Smith, Klonsky, Livingston, & DeMoura, 2004; McGourty, Lewis, Harper, Elkins, & MetzNoseraPapperSanford, 2013). Vineyards, which represent the most common form of row crop agriculture in the catchment, require much less water than other types of crop in Californiadto a depth of as little as 6 cm per area (reported by McGourty et al. [2013] as 0.2 acre-feet of water per acre, which corresponds to 610 cubic meters per hectare)dbut the characteristically poor aquifer characteristics of the regional Franciscan bedrock means that water managers often turn to ambient surface water sources to meet needs. 2.2. Analytical framework To evaluate cumulative impacts of existing water rights and determine whether and where water can be further appropriated within the Navarro River drainage network, we created a GIS-based

16

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

Fig. 1. Navarro River drainage network (by stream order), Mendocino County, CA.

Fig. 2. Streamflow in the Navarro River, median-type water year (1966), as recorded at USGS streamflow gauge 11468000.

catchment-scale model that mirrors the process outlined by the State Water Board for determining water availability in its Policy (SWRCB 2010). According to the Policy, the availability of water for further appropriation is defined by the ratio of the total volume of water among upstream water rights to average discharge over a defined time of yeardin particular, the policy limits new appropriative diversions to a winter diversion season from December 15 to March 31. Water is available for additional appropriation if the ratio of upstream water rights volume to average Policy season

discharge is less than 5 percent at a proposed point of diversion and all points downstream. Our spatially explicit model uses water rights records and streamflow data to calculate the ratio of upstream water rights volume to average discharge for all locations in the drainage network over a defined time of year (such as the Policy diversion season). The model uses a one-third arc-second (approximately 10 m) digital elevation model (obtained from the USGS National Elevation Dataset) as its foundation, and in ArcMap, all existing water rights in the study catchment are added as a point feature shapefile. We populated the water right shapefile with data from the State Water Board's electronic Water Rights Information Management System (eWRIMS), which is a database developed by the state to track water rights information throughout California. Data in eWRIMS describe the location, period of operation, diversion rate, and water right volume for all types of water rights in California, as well as submitted reports from water rights holders describing the actual amount of water used each year (reports of use for riparian and appropriative rights date as far back as 2008). These data are available for water rights throughout the state at http://www.waterboards.ca.gov/waterrights/water_issues/progra ms/ewrims/index.shtml. Approximately two-thirds of the 382 points of diversion from the Navarro drainage network recognized in eWRIMS include a right to divert water during the Policy diversion season (December 15eMarch 31), but almost all allow diversion beyond the established diversion season. To determine availability of water following Policy guidelines, water right volumes for senior water rights during the Policy diversion season are prorated from the total volume of the water right to the diversion season by a ratio of the

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

total number of days of diversion to the number of days in the diversion season. The model then uses the DEM to create a flow accumulation grid for the catchment, and a stream drainage network for all grid cells downstream of water right diversion points. Based on this drainage network, the model calculates the sum of upstream water rights volume during the period of interest (e.g., the Policy diversion season) for each drainage network grid cell. For this Navarro example, diversion point locations were reviewed to ensure they were appropriately located on the DEMderived drainage network. The model also creates an average discharge value for each drainage network grid cell, based on historical data records. For the Navarro River watershed, we used streamflow data from a longterm gauge operated by the US Geological Survey on the Navarro River (Navarro River at Navarro, number 11468000, which operated from 1951 to present, obtained via the USGS website http:// waterdata.usgs.gov/nwis) to calculate average discharge during the defined period of interest. The average discharge value is then scaled to all drainage network grid cells according to a simple drainage basin area-ratio transfer described in the Policy: average discharge at a point in the drainage network is scaled by a ratio of catchment area (i.e., the area-scaled discharge at a point with 1 percent of the upstream area of the Navarro River gauge is 1 percent of the discharge at the Navarro River gauge) and by average annual upstream precipitation (i.e., area-scaled discharge at a point in the drainage network is multiplied by a ratio of average annual rainfall above that point to average annual rainfall above the Navarro River streamflow gauge). Long-term average rainfall data from the PRISM data group (www.prism.oregonstate.edu/) are used as rainfall inputs for scaling streamflow. The Policy assumes that the scaled discharge values represent unimpaired discharge; water is available for appropriation if the sum of upstream water rights is less than 5 percent of the average unimpaired discharge. 2.3. Application The model presented here can be used to evaluate impacts of water rights during time periods beyond the Policy diversion season. To examine how water availability (and inversely, impairment to discharge caused by water rights) changes through the drainage network and over time, we conducted the analysis for four time periods: the Policy diversion season December 15eMarch 31, early winter shoulder season November 1eDecember 14, late winter shoulder season April 1eApril 30, and summer period July 1eJuly 31. The purpose of evaluating water availability during the early and late shoulder seasons is to examine whether water may be available for appropriation during these periods, despite that average flow is likely lower outside of the peak winter rainy season. This may be especially important in the early shoulder season: if the sum of water rights is below the 5 percent threshold for further availability, then it may be a viable option for water users to obtain water during this period and improve water supply certainty. In spring, because of high water needs during April to protect emerging grapes from frost (e.g., Deitch et al. 2009b), appropriations may already exceed desired thresholds for water management. Finally, the purpose of conducting this analysis in summer is to examine the extent to which existing water rights have impaired flow during the time when flow is naturally low and diversions may pose ecological challenges. The availability of water over each period was evaluated under average conditions (as stipulated in the Policy) and under characteristically dry conditions. The purpose of performing this analysis under dry conditions is to examine how much of flow that might occur during each of these study periods could be consumed by water right holders under conditions that are harsher than normal,

17

but still could be expected with regular frequency over the long term. For characteristically dry-type conditions, we chose conditions exceeded by 80 percent of all years (or with a recurrence interval of one in five years) to represent relatively frequent drytype conditions. To illustrate variations in water availability, we classified extent of impairment caused by water rights into five categories: greater than 95 percent of unimpaired discharge (Policy criterion for no further review), between 90 and 95 percent of unimpaired discharge (Policy criterion for requiring additional review), between 80 and 90 percent (unlikely to afford additional appropriation), between 50 and 80 percent (high impact), and less than 50 percent (extreme impact). We also tabulated the length of stream channel subject to each level of impairment for each time period according to stream order. 3. Results The 382 water rights in the Navarro River comprise a total volume of ten million cubic meters (8098 acre-feet); average annual discharge as measured at the Navarro River streamflow gauge is 440 million cubic meters (360,000 acre-feet). In a dry-type year (exceeded 80 percent of all years), Navarro discharge is 45 percent of average discharge (200 million cubic meters). Similarly, dry-type discharge during winter diversion season is 42 percent of average winter diversion season discharge; but dry-type discharge during other periods is much less than average discharge (11 percent of average in November through early December, 25 percent of average in April, and 38 percent of average in July; Fig. 3.) 3.1. Winter season: December 15 to March 31 Calculations of upstream impairment and discharge indicate that there is water available for appropriation during the Policy diversion season throughout most of the Navarro River drainage network. Discharge in the Navarro River and its main tributaries is impaired by less than 5 percent under average flow conditions, even in reaches where water rights are abundant (such as the Navarro River and its tributaries Mill Creek and Anderson Creek; Fig. 4A). Reaches that are heavily impaired are commonly smaller tributaries where one or a few water rights comprise a large amount of the seasonal discharge, but a small amount of the discharge in larger streams such as the Navarro River or other named tributaries. Among streams that are affected by water rights, most of those with discharge impairment greater than 5 percent are zero-, first-, or second-order; though most streams of each stream order are impaired by less than 5 percent (Fig. 5A). Under dry-type conditions, impairment caused by water rights

Fig. 3. Navarro River average monthly discharge and dry-type (exceeded 4 of 5 years) monthly discharge.

Fig. 4. AeB. Impaired discharge caused by water rights, as a percentage of unimpaired flow, through the Navarro River drainage network under average winter Policy Diversion season conditions (top) and dry-type winter Policy Diversion season conditions (bottom).

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

19

Fig. 5. AeH. Impairment to discharge as percent of remaining unimpaired flow caused by water rights, by stream order, under average and dry-type discharge conditions, during the winter Policy diversion season (“Winter”), early winter shoulder season (“NovembereDecember”), late winter shoulder season (“April”), and summer dry season (“July”).

20

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

is more substantial. Most of the tributaries Anderson Creek and Mill Creek are impaired by 5e10 percent by water rights; and impairment is greater than 10 percent in some places, relative to dry-type winter discharge (Fig. 4B). Water rights volumes in other tributaries, such as Indian and Rancheria Creeks, are low compared to dry-type winter discharge. The overall length of zero-, first-, and second-order stream reaches impaired by more than 10 percent is almost double under dry-type conditions, compared to average conditions; but the majority of impaired stream channel is still less than 5 percent by each stream order (Fig. 5B). 3.2. Early shoulder season: November 1 to December 14 Impairment and discharge data during the early winter shoulder season indicate that water is available for appropriation through most of the Navarro drainage during this period. The Navarro River and most of its tributaries are impaired by less than 5 percent under average discharge, and similar to the winter evaluation, heavily impaired reaches tend to be in small streams near areas of abundant water rights (Fig. 6A). More than 80 percent of the streams impaired by more than 10 percent are either zero-, first-, or second-order; all streams fourth-order or greater that are affected by water rights are impaired by less than 5 percent (Fig. 5C). Under dry-type conditions, impairment caused by water rights is more substantial. Almost three-quarters of streams order 0e2 affected by water rights are impaired by more than 10 percent (Fig. 5D), with many of those near Anderson Creek and the Navarro River impaired by more than 20 percent (Fig. 6B). Most of the Navarro and Anderson Creek are impaired at least 10 percent by water rights. Water rights volumes in other tributaries, such as Indian, Mill, and Rancheria Creeks, still are low compared to drytype early winter discharge, resulting in impairment less than 5 percent in these streams. Overall, more small streams are impaired by water rights in the early winter shoulder season, especially under dry conditions. 3.3. Late shoulder season: April 1 to April 30 The dynamics of water availability in the late shoulder season April 1eApril 30 are mostly similar to the early shoulder season. Under average conditions, large streams such as the Navarro and its major tributaries are almost entirely impaired by less than 5 percent; the exception is Anderson Creek, where many water rights (likely for frost protection) impair flow by between 5 and 20 percent (Fig. 7A). The majority of impairment more than 5 percent is among streams of order 0e2, though the majority of each stream order affected by water rights is impaired by less than 5 percent (Fig. 5E). Under dry-type conditions, existing water rights comprise more than 10 percent of discharge throughout the Navarro River (Fig. 7B). Water availability remains variable through major tributaries: discharge in the fifth-order North Fork Navarro River and sixthorder Indian Creek are impaired by less than 5 percent, while discharge in fourth-order Mill Creek and Anderson Creek, and third-order Perry Gulch, are all impaired by more than 10 percent, and upper Rancheria Creek is impaired by more than 5 percent. Overall, more than half of the of streams orders 0e2 affected by water rights are impaired by more than 10 percent; the fraction of streams impaired by more than 10 percent increases by a factor of 3 (Fig. 5F). 3.4. Summer: July 1 to July 31 Approximately 45 percent of streams affected by water rights in

July are impaired by more than 50 percent by existing water rights (Fig. 5G). Even lower portions of Rancheria and Indian Creek, which remained largely unimpaired during the previous analysis periods, are impaired by more than 10 percent (Fig. 8A). Most large-order streams below water rights (e.g., orders 4e7) are impaired more than 10 percent, with most of the Navarro River (7th-order) impaired by more than 50 percent. The headwaters of Indian and Mill Creeks remain only lightly impaired even in summer, even though discharge is low. Impacts of water rights under dry July conditions are more severe than average-type impacts (Fig. 8B): though headwater reaches of Mill, Indian, and Rancheria Creeks remain lightly impaired, the entire Navarro River is impaired more than 50 percent and a larger fraction of the affected drainage network is impaired by more than 20 percent (Fig. 5H). 4. Discussion 4.1. Model applications and limitations Spatial tools play an important role in understanding the effects of development variables that are distributed across a catchment (Xie, Liu, Jones, Higer, & Telis, 2011; Jackson et al. 2013; Castillo, Güneralp, & Güneralp, 2014). Spatially explicit tools allow for characterization of how spatial variables affect water resources over space and time (Perveen & James, 2011; Tong, Sun, Ranatunga, He, & Yang, 2012; Tayyebi, Pijanowski, & Pekin, 2015); in this case, assessment of impairment caused by water rights through the drainage network and comparison of different reaches and subcatchments. The temporal component of our framework is also broadly useful: it allows for assessment of impacts at different times of year which could be set to consider organism life history timing (e.g., salmon spawning periods), or to consider how the cumulative impacts of diversions on streamflow vary under different weather conditions (thus making it useful for climate change scenarios). With large-scale projects (e.g., big dams) posing such substantial challenges to human wellbeing and ecosystem services (Scudder, 2012; Richter et al. 2010), spatial tools will play a critical role for understanding implications of small-scale water management on ecosystems and associated values in the future. The modeling framework presented here has important applications for water resource management beyond this Navarro River case study. Whereas cumulative effects can be, and often are, calculated at a particular point to evaluate impacts of upstream projects, spatially explicit evaluations are useful for regional water resource planning (e.g., Merenlender, Deitch, & Feirer, 2008; Deitch et al. 2013) and for understanding implications of policy scenarios (e.g., Grantham, Mezzatesta, Newburn, & Merenlender, 2013). Evaluations of impacts at particular points can be useful for understanding catchment-wide effects, including variations over time, but they cannot offer insights into variations of impacts over space (Deitch et al. 2009a). The spatially explicit cumulative effects model we present here is a critically important tool for assessing the availability of water, and more broadly, for ensuring adequate ecological protectionsdespecially in watersheds where the cumulative effects of dozens to hundreds of water diversions may result in ecologically significant and spatially variable flow impairment. Developing a computational framework for water rights in California, especially in coastal California where small water rights are abundant and ubiquitous, would represent an important innovation that would benefit aquatic ecosystems as well as water right applicants looking for tools to facilitate the water right application process. Because the framework is in GIS, inputs such as the operation period and volume of particular water rights can be modified, and results quickly re-calculated, to allow analysis of different water management scenarios. Such a model

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

21

Fig. 6. AeB. Impaired discharge caused by water rights, as a percentage of unimpaired flow, through the Navarro River drainage network under average early winter shoulder season conditions (top) and dry-type early winter shoulder season conditions (bottom).

Fig. 7. AeB. Impaired discharge caused by water rights, as a percentage of unimpaired flow, through the Navarro River drainage network under average late winter shoulder season conditions (top) and dry-type late winter shoulder season conditions (bottom).

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

23

Fig. 8. AeB. Impaired discharge caused by water rights, as a percentage of unimpaired flow, through the Navarro River drainage network under average summer (July) conditions (top) and dry-type summer (July) conditions (bottom).

24

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

could be universally applicable any place where water management is decentralized and where ecological protections and senior water right holders need to be considered. Though useful for illustrating the varying availability of water and impacts of water rights through the drainage network, the Navarro River case study presented here is not a perfect means of representing how human water use affects streamflow. Our model incorporates water use as defined in terms of water rights as filed by water users to the State Water Board, and these data are not always perfect representation of the actual volume of water used by water right holders. Water rights data often overestimate the amount of water used at particular locations (as water right applicants aim to maximize their right to water); a recent study showed that claims to water statewide in California greatly exceed the amount of discharge that could occur in absence of humancaused impairments (Grantham & Viers, 2014). Water rights records may also underestimate water uses: most notably, those claiming water under riparian rightdone of the two basic forms of water rights in Californiadhave only recently been required to register their use with the State Water Board. Many water users who have obtained water under riparian right have been slow to adopt this practice (penalties for not reporting were first introduced in 2008), so quantitative information about some riparian rights are not on file and thus were not incorporated into this evaluation. Though not perfect for expressing impairment caused by water management practices, water rights data represent an important synthesis of data for regulatory purposes and an important foundation for understanding the variations in the pressures that human water demands can place on ecosystems over space and time (and are necessary for determining whether water is available for additional appropriative water rights). 4.2. Variations over space and time The above work illustrates how the availability of water and impacts of human water demands on streamflow vary over space and over time. During the recommended diversion season (15 Decembere31 March), the greatest impairments caused by diversions occur on small tributaries (often zero-order, where small reservoirs ranging from 6000 to 120,000 cubic meters accrue water from small unmapped streams), the cumulative effects of which are small compared to discharge from the remainder of the unimpaired drainage network. Low impairment in winter may also be an effect of the seasonal distribution of water rights: though many water rights obtain water in winter (especially recent water rights, to comply with the new instream flow standards), most water requested from the drainage network still occurs during the dry summer (which corresponds with the agricultural growing season). Even along the Navarro River and its tributary Anderson Creek, where most water rights are located, water is available for additional appropriation during the winter diversion season. If five percent impairment to average discharge represents an ecological flow protection standard, then water is available for additional appropriation in shoulder seasons November 1eDecember 14 throughout the region, and April 1eApril 30 through most of the region. Dry-year shoulder season analyses and summer season analyses illustrate substantial cumulative effects of several small diversions on discharge through the largest reaches of the drainage network. Most large-order streams (e.g., order 4e6) affected by water rights are impaired less than 5 percent, and most of the impairment greater than 20 percent is on small streams (order 0e2). However, the abundance of water rights on the Navarro River or on small streams concentrated nearby causes a cumulative reduction of more than 10 percent in the Navarro River during both shoulder

seasons in a dry year, all the way to the ocean. Cumulative effects are more substantial during the summer season, when the Navarro River is impaired by more than 20 percent in an average July and more than 50 percent in a dry July. The differences between average and dry-type hydrologic conditions also influence the cumulative effects of water rights. Drytype discharge (here, periods with discharge exceeded 4 of every 5 years) is less than half of average in each period analyzed in this study; this is because the average is skewed by occasional heavy rains during each shoulder season that do not occur in many years. This multi-annual seasonal variability is a common characteristic of Mediterranean-climate regions (Conacher & Conacher, 1998; Gasith & Resh, 1999). These climatic variations can have profound impacts on aquatic communities (Beche, Mcelravy, & Resh, 2006; Resh et al. 2013) as well as terrestrial ones (Stella, Rodríguez
lez, Dufour, & Bendix, 2013), and low-flow conditions Gonza worsened by instream diversions can further affect aquatic ecosystems (Lawrence, Deitch, & Resh, 2011). Dry-year analyses are especially important for framing climate change discussions, as multi-annual variability is expected to become greater (Micheli, Flint, Flint, Weiss, & Kennedy, 2012) and longer-term droughts are expected through the coming century (Cook, Ault, & Smerdon, 2015). 4.3. Management and ecological implications For prospective applicants looking to acquire appropriative water rights during winter in accordance with ecological protections (streamflow impaired by less than 5 percent), there is water available for diversion during the Policy diversion season, as well as earlier and later in the water year. Because so much water is produced as discharge during the winter rainy season, the threshold of 5 percent of average winter discharge allows most water demands to be met with little impact to winter dischargedeven in a dry year, most major streams are impaired by less than 10 percent. The handful of streams with greater impairment further highlights the importance of spatial tools like this for planning purposes: a few streams are impaired substantially in winter, so additional efforts to improve instream habitat for salmonids in winter may not be beneficial if flows are likely reduced by diversions. The model output can be used to answer many questions that have been raised regarding whether shifting new diversions to winter can protect senior water rights while still providing ecosystem protections. At present, rights to obtain water during the winter Policy diversion season represent less than 2 percent of average seasonal flow throughout the Navarro River, Indian Creek, and Rancheria Creek; and less than 4 percent in Anderson and Mill Creeks. These results indicate that additional water can be appropriated during this season while still maintaining senior water rights and ecosystem protections. These model results also indicate that some portions of the drainage network can support more additional appropriation than others: those streams closer to reaching the 5 percent threshold have less capacity for additional winter appropriation than those with lower percentages of impairment. These spatial variations support the concerns raised by environmental groups about the potential for cumulative impacts, but large impacts are focused on very small streams and are not widespread. The capacity to integrate cumulative effects of water rights over space further highlights the utility of this model for management applications. The results of this study indicate that the overall demand for water in the Navarro River catchment is relatively low, compared to discharge. If the total demand for water, as reflected by water rights, is 11 million cubic meters annually, this water could be

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

obtained from surface water resources during the winter diversion season while still maintaining environmental flows: the average winter discharge is 340 million cubic meters, so all human water needs comprise approximately 3 percent of discharge in the winter Policy diversion season. Even under dry-year conditions, human water needs comprise 8 percent of diversion season discharge. However, the spatial distribution of water rights in winter and at other times of year (e.g., greater in Anderson Creek and Mill Creek than elsewhere in the watershed) indicate that diversion season discharge may not be sufficient to meet all upstream water needs everywhere while still providing adequate ecosystem protections. Despite these challenges, model results from summer months illustrate the importance of reducing diversions in summer: our spatial evaluation indicates that the impairment caused by summer water rights is substantial and extensive throughout the drainage network. Many human demands for water in the Navarro River drainage network are focused in summer, and with low flow due to the seasonality of precipitation, there is not adequate water available to meet human and ecosystem needs. This asynchrony between when people need water (and historically, when they have the right to obtain water) and when it is available is likely a limiting factor to the survival of salmonids in this region (NMFS, 2012). Policies that encourage additional storage in winter to meet summer water demands could be critical in restoring a flow regime suitable for maintaining salmonids and other ecological values in this region. 5. Conclusions Spatially explicit tools are essential for understanding the variations of impacts caused by human activities through the drainage network. They illustrate the spatial arrangement of impacts, and can be used to calculate cumulative effects under several scenarios with varying ecological significance. The iterative capacity of geographic information systems also allows for calculations over different temporal scales. Such tools are especially important for planning purposes, as resource managers are tasked with protecting temporally variable environmental flows over entire catchments relative to decentralized management practices. This research demonstrates that ecologically sustainable water management in the Navarro River may be possible, but will require less water obtained during the summer dry season. The historical practice of obtaining water when needed (during the summer dry season) has resulted in widespread impairment of flow throughout the Navarro River drainage network. Policies that shift the timing of when people can obtain water from summer to winter provides important protections because streamflow is so low during the summer dry season; and the flow standard of 5 percent impairment still is sufficient to provide for all water needs through most of the watershed. Spatial tools such as the one presented here is essential for characterizing the decentralized nature of water management and is necessary for ecologically sustainable water management in the region. Acknowledgments The framework used here was developed with funding from United States Environmental Protection Agency STAR Cumulative Effects Award R829803. The authors also thank two anonymous reviewers for their thoughtful input that improved this manuscript. References Antonino, A. C. D., Hammecker, C., Montenegro, S. L. M. G., Netto, A. M., AnguloJamarillo, R., & Lira, C. A. B. O. (2005). Subirrigation of land bordering small

25

reservoirs in the semi-arid region in the Northeast of Brazil: monitoring and water balance. Agricultural Water Management, 73, 131e147. Arthington, A. H., & Pusey, B. J. (2003). Flow restoration and protection in Australian rivers. Rivers Research and Applications, 19, 377e395. Beche, L. A., Mcelravy, E. P., & Resh, V. H. (2006). Long-term seasonal variation in the biological traits of benthic-macroinvertebrates in two Mediterranean-climate streams in California, USA. Freshwater Biology, 51(1), 56e75. Castillo, C. R., Güneralp, I., & Güneralp, B. (2014). Influence of changes in developed land and precipitation on hydrology of a coastal Texas watershed. Applied Geography, 47, 154e167. Conacher, A. J., & Conacher, J. (1998). Introduction. In Conacher, & M. Sala (Eds.), Land degradation in Mediterranean environments of the world. New York: Wiley. Cook, B. I., Ault, T. R., & Smerdon, J. E. (2015). Unprecedented 21st century drought risk in the American Southwest and Central Plains. Science Advances, 1(1), e1400082. Costanza, R., d'Arge, R., deGroot, R., Farber, S., Grasso, M., Hannon, B., et al. (1997). The value of the world's ecosystem services and natural capital. Nature, 387, 253e260. De Groot, R. S., Wilson, M. A., & Boumans, R. M. (2002). A typology for the classification, description and valuation of ecosystem functions, goods and services. Ecological Economics, 41(3), 393e408. Deitch, M. J., & Kondolf, G. M. (2015). Salmon in a Mediterranean climate: California's incendiary mix. In A. Lassiter (Ed.), Pages 269-291 in sustainable water: Challenges and solutions from California. Berkeley: University of California Press. Deitch, M. J., Kondolf, G. M., & Merenlender, A. M. (2009a). Surface water balance to evaluate the hydrological impacts of small instream diversions and application to the Russian river basin, California, USA. Aquatic Conservation: Marine and Freshwater Ecosystems, 19, 274e284. Deitch, M. J., Kondolf, G. M., & Merenlender, A. M. (2009b). Hydrologic impacts of small-scale instream diversions for frost protection and heat protection in the California wine country. River Research and Applications, 25, 118e134. Deitch, M. J., Merenlender, A. M., & Feirer, S. T. (2013). Cumulative effects of small reservoirs on streamflow in northern coastal California catchments. Water Resources Management, 27, 5101e5118. http://dx.doi.org/10.1007/s11269-0130455-4. Gasith, A., & Resh, V. H. (1999). Streams in Mediterranean climate regions: abiotic influences and biotic responses to predictable seasonal events. Annual Review of Ecology and Systematics, 30, 51e81. Grantham, T. E., Mezzatesta, M., Newburn, D. A., & Merenlender, A. M. (2013). Evaluating tradeoffs between environmental flow protections and agricultural water security. River Research and Applications. http://dx.doi.org/10.1002/ rra.2637. Grantham, T. E., Newburn, D. A., McCarthy, M. A., & Merenlender, A. M. (2012). The role of streamflow and land use in limiting oversummer survival of juvenile steelhead in California streams. Transactions of the American Fisheries Society, 141, 585e598. Grantham, T. E., & Viers, J. H. (2014). 100 years of California's water rights system: patterns, trends and uncertainty. Environmental Research Letters, 9(8), 084012. Habets, F., Philippe, E., Martin, E., David, C. H., & Leseur, F. (2014). Small farm dams: impact on river flows and sustainability in a context of climate change. Hydrology and Earth System Sciences, 18(10), 4207. Jackson, B., Pagella, T., Sinclair, F., Orellana, B., Henshaw, A., Reynolds, B., et al. (2013). Polyscape: a GIS mapping framework providing efficient and spatially explicit landscape-scale valuation of multiple ecosystem services. Landscape and Urban Planning, 112, 74e88. Lawrence, J. E., Deitch, M. J., & Resh, V. H. (2011). Effects of vineyard coverage and extent on benthic macroinvertebrates in streams of northern California. International Journal of Limnology, 47, 347e354. Liebe, J., van de Giesen, N., & Andreini, M. (2005). Estimation of small reservoir storage capacities in a semi-arid environment: a case study in the upper east region of Ghana. Physics and Chemistry of the Earth, 30, 448e454. Malveira, V. T. C., de Arujo, J. C., & Guntner, A. (2011). Hydrological impact of a highdensity reservoir network in Semiarid Northeastern Brazil. Journal of Hydrologic Engineering, 17, 109e117. McGourty, G., Lewis, D. J., Harper, J., Elkins, R., Metz, J., Nosera, J., et al. (2013). Meeting irrigated agriculture water needs in the Navarro river Watershed. Ukiah, California: University of California Cooperative Extension Mendocino County. Merenlender, A. M., Deitch, M. J., & Feirer, S. (2008). Decision support tool seeks to aid stream-flow recovery and enhance water security. California Agriculture, 62, 148e155. Micheli, E., Flint, L., Flint, A., Weiss, S., & Kennedy, M. (2012). Downscaling future climate projections to the watershed scale: a North San Francisco Bay case study. San Francisco Estuary and Watershed Science, 10(4). (NMFS) National Marine Fisheries Service. (2012). Final recovery plan for Central California coast cohosalmon evolutionarily significant unit. Southwest Region, Santa Rosa, California: National Marine Fisheries Service. Perveen, S., & James, L. A. (2011). Scale invariance of water stress and scarcity indicators: facilitating cross-scale comparisons of water resources vulnerability. Applied Geography, 31(1), 321e328. Poff, N. L., Richter, B. D., Arthington, A. H., Bunn, S. E., Naiman, R. J., Kendy, E., et al. (2010). The ecological limits of hydrologic alteration (ELOHA): a new framework for developing regional environmental flow standards. Freshwater Biology, 55, 147e170. Postel, S. L., & Richter, B. D. (2003). Rivers for life: Managing water for people and nature. Washington, D.C., USA: Island Press.

26

M.J. Deitch et al. / Applied Geography 67 (2016) 14e26

^che, L. A., Lawrence, J. E., Mazor, R. D., McElravy, E. P., O'Dowd, A. P., Resh, V. H., Be et al. (2013). Long-term population and community patterns of benthic macroinvertebrates and fishes in Northern California Mediterranean-climate streams. Hydrobiologia, 719(1), 93e118. Richter, B. D. (2010). Re-thinking environmental flows: from allocations and reserves to sustainability boundaries. River Research and Applications, 26(8), 1052e1063. Richter, B. D., Baumgartner, J. V., Powell, J., & Braun, D. P. (1996). A method for assessing hydrologic alteration within ecosystems. Conservation Biology, 1163e1174. Richter, B. D., Davis, M. M., Apse, C., & Konrad, C. (2012). A presumptive standard for environmental flow protection. River Research and Applications, 28(8), 1312e1321. Richter, B. D., Postel, S., Revenga, C., Scudder, T., Lehner, B., Churchill, A., et al. (2010). Lost in development's shadow: the downstream human consequences of dams. Water Alternatives, 3(2), 14e42. Richter, B. D., & Thomas, G. A. (2007). Restoring environmental flows by modifying dam operations. Ecology and Society, 12(1), 12. Richter, B. D., Warner, A. T., Meyer, J. L., & Lutz, K. (2006). A collaborative and adaptive process for developing environmental flow recommendations. River Research and Applications, 22(3), 297e318. Scudder, T. T. (2012). The future of large dams: “Dealing with social, environmental, institutional and political costs”. Taylor & Francis. Smith, R. J., Klonsky, K. M., Livingston, P. L., & DeMoura, R. L. (2004). Sample costs to establish a vineyard and produce wine grapes: North coast region, Sonoma county.

Davis, California USA: University of California Cooperative Extension.
lez, P. M., Dufour, S., & Bendix, J. (2013). Riparian Stella, J. C., Rodríguez-Gonza vegetation research in Mediterranean-climate regions: common patterns, ecological processes, and considerations for management. Hydrobiologia, 719(1), 291e315. (SWRCB) State Water Resources Control Board. (2010). Policy for maintaining instream flows in Northern California coastal streams, SWRCB division of water rights. Sacramento, California USA: California Environmental Protection Agency. Tayyebi, A., Pijanowski, B. C., & Pekin, B. K. (2015). Land use legacies of the Ohio river Basin: using a spatially explicit land use change model to assess past and future impacts on aquatic resources. Applied Geography, 57, 100e111. Tharme, R. E. (2003). A global perspective on environmental flow assessment: emerging trends in the development and application of environmental flow methodologies for rivers. River Research and Applications, 19, 397e441. The Economist. (2007). Australia's water shortage: The big dry (Vol. 383, pp. 81e84). Tong, S. T., Sun, Y., Ranatunga, T., He, J., & Yang, Y. J. (2012). Predicting plausible impacts of sets of climate and land use change scenarios on water resources. Applied Geography, 32(2), 477e489. Wu, J. (2013). Landscape sustainability science: ecosystem services and human well-being in changing landscapes. Landscape Ecology, 28(6), 999e1023. Xie, Z., Liu, Z., Jones, J. W., Higer, A. L., & Telis, P. A. (2011). Landscape unit based digital elevation model development for the freshwater wetlands within the Arthur C. Marshall Loxahatchee national wildlife refuge, Southeastern Florida. Applied Geography, 31(2), 401e412.