RESEARCH ARTICLE availability ratings identify sources for groundwater recharge in the Central The STARR web tool estimates how much and when surplus surface water occurs in each watershed in the three Central Valley basins to help water planners expand groundwater banking.

by Helen E. Dahlke and Ti any N. Kocis

alifornia’s Central Valley produces more than recharge underlying groundwater basins; to make this 400 commodities and 17% of the U.S. total happen, farmers rst need to understand the physical Cagricultural production (valued at nearly $54 distribution and occurrence of excess surface water, billion in 2014) on just 1% of the land in the contiguous particularly the most promising source — high-magni- United States (CDFA 2015). e massive agricultural tude ows. production in the Central Valley has resulted in critical Agriculture in the Central Valley consumes nearly groundwater overdra, triggering state legislation in 40% of California’s annual water supply (surface water, 2014 to require sustainable management of ground- groundwater and reused water developed for agricul- water basins. ere is growing interest in ooding tural, environmental and urban uses), much of it dur- elds during the winter with surplus surface water to ing summer, when surface water supplies are relatively limited (Hanak et al. 2011). Across urban, environmen- tal and agricultural sectors, groundwater accounts for Abstract 38% of the state’s water supply during a normal year, reaching upward of 48% during a dry year (DWR 2015). In California’s semi-arid climate, replenishment of groundwater e constant use of groundwater over the past century relies on precipitation and runo during the winter season. However, has led to a groundwater overdra in the Central Valley climate projections suggest more frequent droughts and fewer years with of over 150 million acre-feet (Faunt 2009). During the above-normal precipitation, which may increase demand on groundwater 2012–2016 severe drought, groundwater depletion aver- resources and the need to recharge groundwater basins. Using historical aged 8.1 million acre-feet per year (Xiao et al. 2017). daily streamow data, we developed a spatial index and rating system of With the passage of the Sustainable Groundwater high-magnitude streamow availability for groundwater recharge, STARR, Management Act (SGMA) in 2014, landowners are in the Central Valley. We found that watersheds with excellent and good now required to implement groundwater sustain- availability of excess surface water are primarily in the Sacramento ability plans by 2040 (SWRCB 2014). One increasingly Basin and northern San Joaquin Valley. STARR is available as a web tool considered approach to achieve groundwater sustain- and can guide water managers on where and when excess surface water ability is managed recharge, which places more is available and, with other web tools, help sustainable groundwater water in groundwater aquifers than would otherwise agencies develop plans to balance water demand and aquifer recharge. naturally occur (Bouwer 2002; Brown and Signor 1974; However, infrastructure is needed to transport the water, and also changes Dillon 2005; Kocis and Dahlke 2017; Scanlon et al. to the legal restrictions on use of such water. 2016). Managed aquifer recharge uses a variety of water

Online: https://doi.org/10.3733/ca.2018a0032

UC Davis researchers developed a web tool that can be used by water managers to identify when — and in which watersheds — excess surface water is available for groundwater recharge.

Helen Dahlke Helen 162 CALIFORNIA AGRICULTURE • VOLUME 72, NUMBER 3 sources (e.g., river water, treated wastewater, desali- nated water) and recharge methods (e.g., inltration basins, injection wells, farmland) to replenish underly- ing aquifers (Dahlke et al. 2018; Dillon 2005; Russo et al. 2015). In recent years, there has been growing interest in ooding farmland with excess surface water in winter to recharge groundwater (Bachand et al. 2014; Dahlke et al. 2018; Kocis and Dahlke 2017). is approach, called on-farm recharge or agricultural groundwater banking, is capable of capturing large volumes of water, particularly from high-magnitude streamows from storm events, which occur frequently during the rainy winter season. However, as groundwater sustainability plans are developed, tools are critically needed to un- derstand the physical distribution and occurrence of such excess surface water ows. Brown Andrew Although aquifer recharge can be conducted with In 2014, California passed any available water (e.g., , recycled water, de- and Dahlke (2017), that could be used by stakeholders legislation mandating the salination, surface water), high-magnitude streamows for improved water resources management. sustainable management of groundwater basins. (i.e., ood ows) likely represent the most accessible One approach for and largest source of water available for future expan- Historical data on high- achieving sustainability, sion of groundwater banking (Harter and Dahlke 2014; magnitude ows called on-farm recharge or Kocis and Dahlke 2017; Scanlon et al 2016). Demand for agricultural groundwater this water during the winter from the agricultural sec- We used historical daily streamow data from United banking, is capable of tor is relatively low. Reservoirs oen make ood control States Geological Survey (USGS) stream gauges in the capturing large volumes releases of stored water in anticipation of large storm Central Valley that had more than 50 years of data (93 of water, particularly from high-magnitude events, and these underutilized releases are expected sites total, visible in gure 1). Gauge sites were classied streamows that occur to increase in frequency and magnitude in the coming as impaired or unimpaired (i.e., unaected by articial during storm events. decades as a result of climate warming (Das et al. 2013; diversions, surface water storage or other works of hu- Dettinger 2011; Hanak and Lund 2012; Yu et al. 2015). mans; Slack et al. 1994) by cross-referencing the sites Additionally, despite overallocation of surface water with the Hydro-Climatic Data Network (HCDN). ir- by the State Water Resources Control Board (SWRCB) teen of the 93 sites were unimpaired. (Grantham and Viers 2014; Scanlon et al. 2016), there is study used the 90th percentile of streamow, seems to exist an abundance of surface water during the calculated from the full record of available streamow winter and wet years, resulting in ood risk for much of data, to designate what constituted a physically avail- the Central Valley (DWR 2016), which could potentially able high-magnitude ow. Most of the stream gauges be reduced by capturing high-magnitude ows. are located downstream of large surface water reser- voirs, thus they are representative of high-magnitude In recent years, there has been ows not captured and stored by surface water reser- growing interest in ooding voirs (e.g., ood releases from reservoirs). Using the 90th percentile was motivated by several factors. First, farmland with excess surface water while most surface water in California is legally allo- in winter to recharge groundwater cated by the SWRCB, high-magnitude streamow (i.e., runo from big storm events including ood ows and (Bachand et al. 2014; Dahlke et al. reservoir releases for ood control) during the winter 2018; Kocis and Dahlke 2017). is oen not. e SWRCB does not currently consider high-magnitude ows for permanent water right/ To evaluate high-magnitude streamow for permit applications or water planning in California, groundwater recharge eorts, we conducted a statisti- but estimates surface water availability purely on long- cal analysis of historical daily streamow records to term average ow, which ignores a large fraction of the provide insights into its physical availability and spatial streamow available during the winter rainy season. distribution (Kocis and Dahlke 2017). Our goal was to Second, the 90th percentile is oen used by the USGS create an index and web application of high-magnitude and the environmental ow community to designate streamow availability for groundwater recharge across ows as “much above normal,” or as “high” (Henriksen the Central Valley, including a relative comparison of et al. 2006; Olden and Po 2003; Richards 1990; USGS watersheds for dierent time periods (e.g., November to 2016). Hence, we adopted this threshold for our desig- April), using the statistical analyses presented in Kocis nation of high-magnitude streamow.

http://calag.ucanr.edu • JULY SEPTEMBER 2018 163 calculated over two periods of record: (1) the full record of available data and (2) the record of data since the most recent impairment (i.e., post-impairment period). For the Basin and the San Joaquin Valley (comprised of the Basin and the Tulare Lake Basin), the post-impairment periods of record are 1970 to 2014 and 1989 to 2014, respectively. Flow metrics were also calculated for dierent time periods and the ve water year types (critical, dry, below normal, above normal and wet) dened in the San Joaquin Valley Index and Sacramento Valley Index (SWRCB 1995, Null and Viers 2013). Time periods were the hydrologic year (Oct. 1 to Sept. 30), the winter rainy season (November to April), the winter season (December to February) and each month between November and April. Stream ow availability ratings

We developed a streamow availability rating for re- charge (STARR) based on three of the ve high-mag- nitude streamow metrics that emerged to be the most indicative for the availability of excess surface water from high-magnitude ow: magnitude, duration and interannual frequency. STARR was calculated using an empirical weighting method, the rank-ordered centroid weighting method (Barron and Barrett 1996), that combines the weighted, ranked equal-area ratings of the magnitude, duration

USGS California Water Science Center Water USGS California and interannual frequency metrics for the contributing U.S. Geological Survey watershed of each stream gauge. For the watersheds of technicians prepare to Stream ow metrics the 93 stream gauges analyzed by Kocis and Dahlke go out to the center of For each stream gauge, ve statistical metrics were cal- (2017), STARR was determined for three time periods: the in the San culated to inform on the availability of high-magnitude the hydrologic year, November to April, and December Joaquin River below Friant ows: magnitude, duration, timing, intra-annual fre- to February. to get an accurate quency and interannual frequency (Kocis and Dahlke measurement of water 2017). Magnitude is the total ow volume above the Watershed scale ow during releases from the dam for ood control. 90th percentile. Duration is the number of days above e STARR index displays the high-magnitude ow Capturing such releases the 90th percentile. Timing is the day of the hydrologic availability (developed from point gauge locations) for for aquifer recharge could year (DOHY) of the center of mass (COM) of ows groundwater recharge at the watershed scale. e up- potentially reduce winter above the 90th percentile; COM is dened by the day stream contributing watershed was derived for each of ood risk in the Central when 50% of the total ow volume above the 90th per- the 93 stream gauges using the 1-arc-second National Valley. centile in a given time period (e.g., winter) has passed. Elevation Dataset (USGS 2009). Watersheds were de- Intra-annual frequency is the count of 1-day peaks lineated in ArcGIS with a drainage area threshold (ow that occur over the 90th percentile; a 1-day peak occurs accumulation threshold) equal to 1% of the maximum on a day when the ow is higher than both the previous ow accumulation. If several stream gauges exist along day and the next day. Interannual frequency is the frac- the same river, the contributing areas were visually tion of years with ow above the 90th percentile. Not overlaid such that smaller upstream watersheds remain all years in the historical records have ow above the on top of larger downstream watersheds. For accuracy, 90th percentile; therefore, these years were excluded the upstream drainage areas were compared to those from the calculation of the magnitude, timing, dura- provided by USGS. tion and intra-annual frequency metrics. is process is referred to as zero deation. Four-step process e STARR calculation process included four steps: Calculation periods First, three spatial ow metrics, V/A, D/P and YWF, All metrics reect the average value over specic cal- were developed for the watershed of each of the 93 culation periods. Given potential changes in the ow stream gauges. e magnitude, V (in cubic kilometers), regime due to the construction of or diversions was standardized by the watershed area, A (square within the watershed of each gauge, the metrics were kilometers). e duration, D, was standardized by the

164 CALIFORNIA AGRICULTURE • VOLUME 72, NUMBER 3 number of days in the calculation period, P (e.g., 90 STARR values (ranging from 1 to 6) were then split days from December to February), and the interannual into six equal intervals corresponding to the STARR frequency (fraction of years with high-magnitude ow) classes (e.g., 1 < = numeric STARR < 1.83 is very poor). was le unchanged and notated as YWF. e numerical STARR values were categorized as very en, using an equal-area classication method, the poor, poor, moderately poor, moderately good, good spatial ow metrics were scored from 1 to 6. First, val- and excellent. ues in each metric were sorted from smallest to largest. Watersheds with a high STARR (i.e., excellent rat- Next, the score was determined relative to other water- ing) are generally characterized by large high-magni- sheds in the study region, the Central Valley, by break- tude ows that persist for a high number of days and ing the metric values into six equal-area classes with the recur, on an interannual basis, with high frequency. same number of watersheds in each class. Conversely, watersheds with a low STARR (very poor Next, the rank-ordered centroid method (Barron rating) generally contribute less high-magnitude ow and Barrett 1996) was used to develop an additive than other watersheds within the study area, and the multi-attribute model for the three spatial ow metrics ow occurs only for a very short period of time each to calculate the nal numeric STARR value. e model year, and recurs, on an interannual basis, with low fre- allows developing an empirical formula to calculate quency. A high STARR value corresponds to watersheds the STARR and determining weighting coecients for that have excellent physical surface water availability the three spatial ow metrics considered in the for- relative to other watersheds in the Central Valley, while mula based on their relative importance as outlined in a low STARR value corresponds to watersheds that have Barron and Barrett 1996. In this formula (see equation very poor physical surface water availability. (1) below), the interannual frequency was given the highest weight (considered the most important factor), Optimal month rating duration was given the lowest weight (considered the least important factor, of the three) and the weight on We used the STARR index to also develop a decision magnitude was chosen to fall in between. e resulting support tool for the optimal time when high-magnitude FIG. 1. Streamow STARR formula was streamow is available for recharge across the Central availability rating for STARR = 0.611 × ranked(YWF) + 0.277 × Valley. We developed the optimal month rating (OMR) recharge (STARR) for the ranked(V/A) + 0.111 × ranked(D/P) (1) to answer the question: “For any watershed, what is the Sacramento River Basin, month in which most high-magnitude ow is physically San Joaquin River Basin e nal STARR ratings were calculated by entering available?” e OMR identies the month within a time and Tulare Lake Basin during December to the rst round of values from step 2 into equation (1). period of interest that provides the greatest water avail- February (A), November Using an equal-interval ranking, the resulting numeric ability and highest ow reliability (longest duration, to April (B) and the hydrologic year (C).

STARR (Dec – Feb) STARR (Nov – Apr) STARR (Hydrol. year) (A) Excellent (B) Excellent (C) Excellent Good Good Good Moderately good Moderately good Moderately good Moderately poor Moderately poor Moderately poor Poor Poor Poor Very poor Very poor Very poor No rating No rating No rating and Streams and canals Streams and canals Aqueduct Aqueduct Canal Canal Streams Streams Streams Sites Sites Sites

0 20 40 60 Miles 0 20 40 60 Miles 0 20 40 60 Miles

http://calag.ucanr.edu • JULY SEPTEMBER 2018 165 TABLE 1. Spatial distribution of streamow availability ratings for recharge (STARR) for the contributing areas of 93 stream gauges within the Sacramento River Basin and San Joaquin Valley

December to February November to April Hydrologic year STARR Contributing area Contributing area Contributing area

acres # sites acres # sites acres # sites Excellent 2,022,829 13 2,140,252 13 109,142 3 Good 4,295,153 17 5,470,998 18 2,664,422 22 Moderately good 11,042,192 22 9,360,678 16 4,401,728 20 Moderately poor 12,846,026 12 2,039,866 15 2,839,750 21 Poor 2,164,866 15 2,435,685 17 12,586,170 19 Very poor 3,032,128 14 13,955,715 14 12,801,980 8

greatest frequency and/or largest volume) and is deter- unimpaired from the Coast Ranges and the mined by estimating the highest monthly STARR value Sierra Nevada (g. 1A). Watersheds with poor or very for each watershed in a given time period (e.g., Novem- poor December to February STARR are primarily High ber to April). Sierra watersheds, where streamow availability dur- OMR is determined in a similar fashion to STARR, ing winter is limited because most precipitation falls but instead of comparing one watershed to another as snow and only becomes snowmelt runo during the watershed, OMR compares the water availability in a and early summer. In contrast, the Tulare Lake single watershed from one month to another month Basin is ranked moderately poor. Here, streamow within a given time period. OMR is determined for availability is limited, occurring only during winter two time periods: December to February and November (December to February) when rainfall produces runo, to April. or ooding during wet years. For the November to April period, the distribution Stream ow availability for recharge of watersheds with excellent and good streamow avail- highest in Sacramento River Basin ability for recharge (g. 1B, table 1) is similar to that for the December to February period. Most of those wa- STARR was calculated for the Sacramento River Basin, tersheds are in the Sacramento River Basin, occupying Between the San Joaquin River Basin and the Tulare Lake Basin about 42% of the basin. Streamow availability for re- December — an area of over 35.4 million acres (about one-third of charge in the San Joaquin River Basin and Tulare Lake the land surface of California) (g. 1). Kocis and Dahlke Basin is predominately moderately poor to very poor, and February, (2017) estimated that in an average year with high- with a few moderately good ratings on the west side of 30 of the 93 magnitude ow (e.g., years when streamow exceeds the San Joaquin River. Most of the valley oor of the watersheds the 90th percentile ow), between November and April Tulare Lake Basin has very poor streamow availability approximately 1.88 million acre-feet (2.31 cubic kilome- for groundwater recharge. have excellent ters) and 0.97 million acre-feet (1.2 cubic kilometers) of During the hydrologic year (Oct. 1 to Sept. 30), or good high-magnitude ow are exported from the Sacramento streamow availability for groundwater recharge River Basin (USGS site 11447650) and the San Joaquin varies between good and very poor throughout the stream ow River Basin (USGS site 11303500), respectively, to the Central Valley (g. 1C). e Tulare Lake Basin gen- availability, Sacramento–San Joaquin Delta. erally has very poor excess streamow availability; an area of Most of the watersheds with the greatest high- high-magnitude streamow availability in the largest magnitude ow availability are in the Sacramento River tributaries to the San Joaquin River from the Sierra over 6.3 Basin, which generally receives more precipitation than Nevada is either poor or moderately poor. e low million acres. the southern Central Valley. e interannual frequency ratings can be attributed to the low interannual fre- indicated that those watersheds carry high-magnitude quency at which high-magnitude ows occur and the ows on average in 7 to 9 out of 10 years, while water- low overall magnitude of those ows compared to the sheds in the San Joaquin River Basin and Tulare Lake size of the watersheds. Basin carry high-magnitude ows on average in only 2 Within the Sacramento River Basin during the to 5 out of 10 years (Kocis and Dahlke 2017). hydrologic year, most watersheds below major rim res- STARR shows distinct patterns for the high- ervoirs, including the upper Sacramento and Feather magnitude ow availability in winter and during the , are rated poor or moderately poor, while most hydrologic year (g. 1; table 1). Between December and unimpaired streams and headwater catchments are February, 30 of the 93 watersheds have excellent or rated good or moderately good. Unimpaired watersheds good streamow availability, an area of over 6.3 million in the Sacramento River Basin lack major control struc- acres, or 17.8% of the study area (table 1). ese water- tures (e.g., surface reservoirs, major diversions) and sheds are primarily in the Sacramento River Valley and therefore carry high-magnitude ow for longer periods the northern San Joaquin Valley and include mainly throughout the winter and, in some cases, the shoulder

166 CALIFORNIA AGRICULTURE • VOLUME 72, NUMBER 3 months (September to November, March to July), Our STARR web tool provides another layer of in- which results in good and moderately good STARR for formation that growers, landowners and water district the hydrologic year. For most of the impaired streams managers can use to plan agricultural groundwater in the Sacramento River Basin, the high-magnitude banking programs as part of their operations or to ow availability in spring, summer and fall is regulated develop groundwater sustainability plans. We envision or prevented by major reservoirs. the STARR web tool being used at a local scale with other tools. For example, to identify suitable recharge Optimal months for recharge areas within groundwater management areas, SAGBI identied could be used to identify areas with suitable soils, land use surveys could be used to identify areas with In the Sacramento River Basin, the OMR indicates suitable crops, the eWRIMS (Electronic Water Rights that for the December to February period, February is Information Management System) database could be the ideal month for groundwater recharge in terms of used to identify current surface water right holders, streamow availability for recharge (g. 2). In the San and infrastructure maps could be used to identify ar- Joaquin River Basin and Tulare Lake Basin, the ideal eas with the necessary conveyance infrastructure. e month for groundwater recharge is January for the intersection of soil suitability, crop suitability, land use, December to February period, and December for some conveyance infrastructure, legal water availability, and of the Sierra Nevada watersheds. During December to physical water availability (STARR) is the ideal loca- February, water availability stems primarily from rain- tion to conduct an agricultural groundwater banking fall, not snowmelt. project. Excess surface water availability that includes snow- A recent survey conducted by the Public Policy melt is reected in the November to April OMR (g. Institute of California indicates that many water dis- 2B), which identies March as the ideal month for the trict managers cite infrastructure capacity constraints San Joaquin River Basin and Tulare Lake Basin, and as a major barrier to the expansion of existing recharge February, again, as the ideal month for the Sacramento programs (Hanak et al. 2018); many water districts River Basin. In the Sierra Nevada tributaries, April is and counties currently lack the infrastructure to divert the ideal month for streamow availability, clearly in- high-magnitude ows (DWR 2017; O’Geen et al. 2015). dicating snowmelt as the main source. e STARR web tool can be used to advise the plan- ning and expansion of water conveyance systems at Web tool for decision-making district and watershed scales to deliver water to suitable recharge lands, since it provides detailed information FIG. 2. Optimal month rating (OMR) for e streamow availability metrics (magnitude, tim- on the volumes and timing of when high ows are groundwater recharge ing, duration, interannual and intra-annual frequency available. Comparing peak ow volumes to the capacity based on high-magnitude of ow above the 90th percentile) and STARR and of existing conveyance systems allows water managers streamow availability in OMR maps are available as an interactive web tool (g. December to February (A) 3) at http://recharge.ucdavis.edu/STARR. More infor- and November to April (B). mation on it is provided in the technical appendix. OMR (Dec – Feb) OMR (Nov – Apr) Web-based decision support tools have proven use- (A) December (B) December ful in supporting farmers and landowners in decision- January January making processes. For agricultural groundwater - February February ing, several such tools help identify suitable on-farm No rating March Streams and canals recharge locations. For example, the Soil Agricultural April Aqueduct No rating Groundwater Banking Index (SAGBI) (casoilresource. Canal Streams and canals lawr.ucdavis.edu/sagbi/) developed by O’Geen et al. Streams Aqueduct (2015) recommends recharge locations based on soil Sites Canal suitability by considering ve factors: soil prole Streams percolation rate, root zone residence time, chemical Sites limitations, topography and soil surface conditions. Crop suitability research for groundwater recharge is available for alfalfa (Dahlke et al. 2018) and almonds (Volder et al. 2016). e California Department of Water Resources recently published a detailed land use survey of cropland parcels within California, distin- guishing 12 dierent land use classes (gis.water.ca.gov/ app/CADWRLandUseViewer/). is land use survey represents the most accurate (i.e., in terms of cropping system and spatial accuracy) land use and land cover dataset available to date for water resources planners for water budget calculations.

http://calag.ucanr.edu • JULY SEPTEMBER 2018 167 FIG. 3. Interactive, web- based map displaying the magnitude, timing, duration, interannual and intra-annual frequency of high-magnitude ows for the 93 sites analyzed within the Central Valley. Users can click on a site of interest on the map, which loads a graph for the selected high- magnitude streamow metric, record length and time period. In the example shown above, users can see the average high-magnitude ow volume for the stream gauge at the Mokelumne River below Camanche Dam for each month between November and April and each water year type. The displayed ow volumes are based on estimates considering the full record of streamow data available for that site. to determine the need for expansion or new construc- transfers would allow local groundwater sustainability tion of conveyance infrastructure. agencies to address many of the undesirable results listed in SGMA, including land subsidence, lowering of Transfers from north to south the water table and groundwater storage depletion. Additionally, knowledge that in the San Joaquin As indicated in gure 1, streamow availability for River Basin and Tulare Lake Basin most high-mag- groundwater recharge in the Central Valley displays a nitude ow originates from spring snowmelt could clear north-south gradient, with highest availability in motivate changes in reservoir operation policies; more the Sacramento River Basin. e availability of excess water from reservoirs could be released during late surface water is spatially disparate to agricultural land fall or early winter for groundwater recharge to cre- suitable for groundwater recharge. e most highly ate increased exibility for ood management of more rated soils (with a SAGBI rating of excellent, good or intense precipitation events (Hanak and Lund 2012; moderately good) are located on the broad alluvial fans Maurer and Duy 2005). on the east side of the Central Valley (near the Moke- lumne, Stanislaus, Merced, Kern and Kings rivers); in Water rights, legal use of ows the Sacramento Valley, suitable land is limited to nar- row alluvial bands along the Sacramento River and its Given that groundwater recharge is not considered a major tributaries. “benecial use” in the California Water Code (Cali- is spatial disparity in water availability and suit- fornia Water Code 2017), the legal use of the high- able recharge locations seen in nearly every visualiza- magnitude ows calculated in this study remains tion of STARR suggests long-term investment is needed questionable for the near future. Landowners and water in statewide water projects that could convey excess districts planning new groundwater recharge programs surface water from the Sacramento River Basin to suit- will likely have to obtain a new surface water right or able land in the southern Central Valley. Water trans- change an existing water right. fers would allow balancing the critically overdraed e SWRCB currently estimates the water avail- Kern, Kings, Tule, Kaweah, and Tulare Lake ground- ability for a new appropriative surface water right water basins, where there is a generally high water de- using a method similar to the rational runo method mand (DWR 2016). (Kuichling 1889; SWRCB 2001), which estimates the OMR indicates that water transfers from the average annual unimpaired runo at a diversion point Sacramento River Basin, where high-magnitude ow of interest considering only the contributing area, aver- is available as early as November, to the San Joaquin age annual precipitation and the land use within the and Tulare Lake Basins could be made early in the watershed (SWRCB 2001). is conservative method winter, when crops such as alfalfa and almonds are still is used to ensure that there is “unappropriated water dormant and before local ood ows from the High available to supply the applicant” (Water Code section Sierra become available in March or April. Such water 1375(d)), while accounting for “the amounts of water

168 CALIFORNIA AGRICULTURE • VOLUME 72, NUMBER 3 needed to remain in the source for protection of bene- cial uses” (Water Code section 1243), such as recreation References and the preservation of sh and wildlife habitat. Bachand PAM, Roy SB, Choper- Grantham TE, Viers JH. 2014. 100 Scanlon BR, Reedy RC, Faunt CC, ena J, et al. 2014. Implications of years of California’s water rights et al. 2016. Enhancing drought However, as indicated by Grantham and Viers using on-farm ood ow cap- system: Patterns, trends and resilience with conjunctive use (2014), in many areas of California, mainly the Central ture to recharge groundwater uncertainty. Environ Res Lett and managed aquifer recharge and mitigate ood risks along 9(8):084012. in California and Arizona. Envi- Valley, surface water has been overallocated to the the Kings River, CA. Environ Sci Hanak E, Lund JR. 2012. Adapt- ron Res Lett 11(3):035013. extent that surface water rights account for nearly 10 Technol 48(23):13601–9. ing California’s water manage- Slack JR, Lumb AM, Landwehr times the natural surface water supplies. is, theoreti- Barron FH, Barrett BE. 1996. ment to climate change. Clim JM, US Geological Survey. 1994. Decision quality using ranked Change 111:17–44. Hydro-Climatic Data Network cally, precludes any additional appropriation of surface attribute weights. Manage Sci Hanak E, Lund J, Dinar A, et al. (HCDN)—A USGS Streamow water. Yet, over-appropriation is, to a large extent, an 42(11):1515–23. 2011. Managing California’s Wa- Data Set for the US for the artifact of the water availability analysis conducted by Bouwer H. 2002. Articial ter: From Conict to Reconcilia- Study of Climate Fluctuations. recharge of groundwater: Hy- tion. San Francisco: Public Policy US Department of the Interior, the SWRCB, which is based on average annual ows drogeology and engineering. Institute of California. 482 p. US Geological Survey. and does not take into account the large variability in Hydrogeol J 10(1):121–42. Hanak E. 2018. The State of [SWRCB] State Water Resources streamow. Hence, new permitting approaches that Brown RF, Signor DC. 1974. Ar- Groundwater Recharge in the Control Board. 1995. Water ticial recharge — state of the San Joaquin Valley. www.ppic. Quality Control Plan for the San would legally permit the use of high-magnitude ow for art. Groundwater 12(3):152–60. org/blog/state-groundwater- Francisco Bay/Sacramento- groundwater recharge are needed. recharge-san-joaquin-valley/ San Joaquin Delta . California Water Code. 2017. 95-1WR. www.waterboards. Allowing a water right permit for the diversion http://leginfo.legislature.ca.gov/ Harter T, Dahlke HE. 2014. Out ca.gov/waterrights/water_is- of “high ow” could bridge the gap between policy faces/codesTOCSelected.xhtml? of sight but not out of mind: sues/programs/bay_delta/ tocCode=WAT&tocTitle=+Water California refocuses on ground- wq_control_plans/1995wqcp/ requirements (such as the need for a temporary or per- +Code+-+WAT water. Calif Agr 68:54–5. docs/1995wqcpb.pdf (accessed manent water right for surface water diversions), legal [CDFA ] California Depart- Henriksen JA, Heasley J, Kennen Aug. 3, 2018). requirements (for stream reaches that are already legally ment of Food and Agriculture. JG, Nieswand S. 2006. Users’ SWRCB. 2001. Evaluation of 2015. California Agricultural Manual for the Hydroecological State Water Resources Control over-appropriated) and physical surface water availabil- Statistics Review, 2015-2016. Integrity Assessment Process Board Water Availability Analy- ity for groundwater recharge (in the form of ood ows www.cdfa.ca.gov/statistics/ Software (Including the New sis. MBK Engineers report. 8 p. PDFs/2016Report.pdf Jersey Assessment Tools). Res- during above normal or wet years). Such permits would www.cwemf.org/workshops/ Dahlke HE, Brown A, Orlo S, ton, VA: US Geological Survey. mbkreport.pdf have to agree on legally acceptable high ow thresholds et al. 2018. Managed winter Kocis TN, Dahlke HE. 2017. SWRCB. 2014. Sustainable at the point of diversion to ensure that high ow diver- ooding of alfalfa recharges Availability of high-magnitude Groundwater Management groundwater with minimal streamow for groundwater sions for groundwater recharge do not cause injury to Act §§ 346-1-10, §§ 347- crop damage. Calif Agr 72(1):65- banking in the Central Valley, 1-23, §§ 348-1-3. State of existing water rights holders or environmental ow 75. https://doi.org/10.3733/ California. Environ Res Lett California. www.water.ca.gov/ considerations. ca.2018a0001 12(8):084009. cagroundwater/docs/2014%20 Das T, Maurer EP, Pierce DW, Kuichling E. 1889. The relation Sustainable%20Groundwa- Permits could be restricted to the winter period, et al. 2013. Increases in ood between the rainfall and the ter%20Management%20Leg- from November to March, and dene strict in-stream magnitudes in California under of sewers in popu- islation%20with%202015%20 amends%201-15-2016.pdf ow requirements (e.g., the passage of channel-forming warming climates. J Hydrol lous districts. Trans Am Soc Civil 501:101–10. Eng 20:1–56. [USGS] United States Geological ows or fall ushing ows for and nutrient Dettinger M. 2011. Climate Maurer EP, Du y PB. 2005. Survey. 2009. National Elevation transport) (see Kocis and Dahlke 2017 for a more de- change, atmospheric rivers, Uncertainty in projections of Dataset (NED). US Department tailed discussion of these considerations). Solving these and oods in California – a streamow changes due to of the Interior, US Geological multimodel analysis of storm climate change in California. Survey. http://nationalmap.gov/ regulatory challenges to groundwater recharge, along frequency and magnitude Geophys Res Lett 32:1–5. elevation.html with a potential expansion of infrastructure capac- changes. J Am Water Resour As Null SE, Viers JH. 2013 In bad USGS. 2016. Daily Streamow 47:514–23. ity, will open new avenues to greater water security in waters: water year classication Conditions, USGS Current c Dillon P. 2005. Future manage- in nonstationary climates. Water Water Data for the Nation. US California. ment of aquifer recharge. Hy- Resour Res 49:1137–48. Department of the Interior, US Geological Survey. http://water- drogeol J 13(1):313–6. O’Geen AT, Saal MBB, Dahlke datausgs.gov/nwis/rt [DWR] California Department HE, et al. 2015. Soil suitability H.E. Dahlke is Associate Professor in Integrated Hydrologic Sciences of Water Resources. 2015. index identies potential areas Volder A, Shackel K, Lamp- California’s Most Signicant for groundwater banking on inen B, et al. 2016. Winter and T.N. Kocis is Ph.D. Student in the Hydrologic Sciences Graduate Droughts: Comparing His- agricultural lands. Calif Agr Recharge and Irrigation. Group, Department of Land, Air and Water Resources, UC Davis. torical and Recent Conditions. 69(2):75–84. UC Davis Dept of Plant Sci- ences. https://lf.almonds. Sacramento: DWR. www. Olden JD, Po NL. 2003. Redun- com/WebLink/ElectronicFile. water.ca.gov/waterconditions/ dancy and the choice of hydro- aspx?docid=2333&dbid=0 docs/California_Signcant_ logic indices for characterizing Droughts_2015_small.pdf streamow regimes. River Res Xiao M, Koppa A, Mekon- DWR. 2016. Groundwater Basins Appl 19:101–21. nen Z, et al. 2017. How much groundwater did California’s Subject to Critical Conditions of Richards RP. 1990. Measures of Central Valley lose during the Overdraft. www.water.ca.gov/ ow variability and a new ow- 2012–2016 drought? Geophys groundwater/sgm/pdfs/COD- based classication of Great Res Lett 44(10):1–8. basins_2016_Dec19.pdf Lakes tributaries. J Great Lakes DWR. 2017. Water Available for Res 16:53–70. Yu Z, Jiang P, Gautam MR, et al. 2015. Changes of seasonal Replenishment. http://wdl.wa- Russo TA, Fisher AT, Lockwood storm properties in California ter.ca.gov/groundwater/sgm/ BS. 2015. Assessment of and Nevada from an ensemble wafr.cfm managed aquifer recharge of climate projections. J Geo- Faunt CC. 2009. Groundwater site suitability using a GIS phys Res-Atmos 120:2676–88. Availability of the Central Valley and modeling. Groundwater Aquifer, California. USGS profes- 53(3):389–400. sional paper No. 1766. Sacra- mento: US Department of the Interior, US Geological Survey.

http://calag.ucanr.edu • JULY SEPTEMBER 2018 169