Central Valley Surfacewater Map

About this map

Every line and circle is tied to a published, measured record. Where no public record exists, the feature is drawn dashed grey, and nothing is filled in. The exceptions are numbers modelled by this project or by PPIC: the estimated losses block in the System accounting panel (and the same estimates in the Pick a stretch of waterway panel), the “Estimate” rows for the Sacramento Slough and Colusa Basin Drain inflows under the Wilkins Slough to Verona balance and under a picked stretch that includes them, the flows the stretch budget works out between two gauges, and PPIC's modelled split of Delta outflow for WY2001–2021. All are kept in separate files, headed “Estimate” or “PPIC estimate”, and never mixed into the measured flows or balances (see “Estimated losses” below). The slough has no record after February 2013, and the drain none after September 2012 (a stand-in to September 2020, and none in a month with too few days), so their rows give a range from their own measured years. That range is never added to the balance. A few published figures are calculated, not measured: DWR's in-Delta use and net Delta outflow, and DWR's 2023 modelled aqueduct capacity. They are labelled estimated or modelled where they appear.

The one forecast is the Allocation outlook: a modelled range for a contract's final allocation, fitted on the published announcements. It opens in its own window, is labelled “Forecast: modelled, not measured” wherever it appears, and is never added to a measured flow or balance.

Delivered project water, in the Surface water by subbasin / GSA section, shows the project water delivered to the contractors whose districts lie wholly inside each subbasin or GSA: a map view that colours every subbasin or GSA for the year of the time bar (off by default; every other river, canal and reservoir is hidden while it is on), a chart over the years, a ranked table, and acre-feet per acre of the whole boundary area. Only the published figures of those contractors colour, rank and chart; contractors that straddle a boundary are shown beside them, marked ~, as an area-share estimate. CVP is per water year and SWP per calendar year, and the two are never added. A year with an unknown figure is n/a, never zero. Local surface water is not included.

Sources

• CDEC (California Data Exchange Center, DWR): reservoir storage, inflow and release; SWP and CVP pumping plants (Banks, Jones, Barker Slough, South Bay, Gianelli, Dos Amigos, Las Perillas, Buena Vista, Edmonston, Oso); Friant and Goodwin canal releases; river gauges; DWR's Delta outflow estimate.
• DWR Dayflow: the daily Delta water balance (inflows, exports, in-Delta use, net outflow), WY1997 through the latest published water year. After that the map uses CDEC equivalents and, for net outflow and in-Delta use, Reclamation Central Valley Operations' daily Delta Outflow Computation (the projects' own accounting). In-Delta use (net of rain on the Delta) and net outflow are calculated, not measured, so they are marked estimated; the depletion behind both is a fixed 1950s-60s curve. Dayflow's WY2023 has no Mokelumne River inflow; the map adds the USGS Woodbridge record (see Data notes).
• Revised Water Right Decision 1641 (2000), Table 3 and Table 4, and DWR Bulletin 120 (February–May issues, WY2001 on): the Delta flow requirements as written, encoded number by number with their table, page and footnote. Also encoded are the preliminary year types and previous-month Eight River Index that DWR published, which set several of them. Shown in the D-1641 compliance indicators section and in Year in review, beside the records on this map. The requirement is never mixed into a measured balance. Salinity (CDEC daily mean EC at Collinsville, Mallard Island and Port Chicago) evaluates the standard's salinity alternatives where a station record is available. The salinity objectives themselves are in Tables 1–3: municipal chloride, agricultural EC at the western, interior and southern Delta stations and the export area, San Joaquin River EC for striped bass, and Suisun Marsh. They are checked against CDEC's EC and chloride records, at the stations the Projects report or at a labelled nearby station. We pick the Suisun Marsh high-tide EC from CDEC's hourly EC and stage.
• PPIC Delta Water Accounting (Gartrell, Mount, Hanak, Sencan, Public Policy Institute of California, 2022): PPIC's modelled split of Delta outflow into what the standards required, what export limits left in the Delta and what was uncaptured. It also gives two other estimates of in-Delta use (DWR DICU, DCD) and watershed totals. All cover WY2001–2021. Shown only in blocks labelled “PPIC estimate”, never in the measured balances. Data set.
• DWR SWP Monthly Reports of Operations: daily pumping at Banks, Barker Slough, South Bay, Dos Amigos, Gianelli (pumping and generation), Las Perillas, Buena Vista and Edmonston, from the report PDFs DWR once posted (now kept by the Internet Archive). They fill the aqueduct record before these plants appear on CDEC. Each parsed month is checked against the report's own monthly totals.
• USGS NWIS, Delta channels: tidally filtered daily flows on the Delta Cross Channel, Georgiana Slough, Old River at Bacon Island, and Middle River, the channels that carry Sacramento River water through the Delta to the export pumps. Shown under “Channel flows” in the Delta panel, signed, and kept out of the balance totals.
• State Water Board eWRIMS and CalWATRS (self-reported): the annual reports of diversion and use that water-right holders file (eWRIMS to report year 2024, CalWATRS from 2025), summed to water years. They are the holders' own claims, not measurements, so they appear only in a separate card headed “Self-reported to the State Water Board” under a selected diversion. The card sits beside the measured volume, and the filings are never added to it or to any balance. A same-source badge marks the districts whose filing and the delivery table agree to within 2 acre-feet a month (one set of records, not two checks). Storage rights, project water, overlapping rights and settlement-contract years are labelled where they apply. Rights are matched to diversions by holder name and place; weak matches are listed, not compared.
• USGS NWIS: daily flows for irrigation canals (Modesto, Turlock, Woodbridge, Feather River settlement canals, and Putah South) and several rivers. Some are published as approved data only, about a year behind.
• Reclamation CVO monthly delivery tables, which appear at the monthly and water-year steps only:
  • Table 21: canal volumes where no daily record exists (Tehama-Colusa and Corning canals; Folsom South before 2016).
  • Tables 22 (Friant-Kern Canal, with the districts' pump-ins) and 23 (Madera Canal): the Friant Division's deliveries.
  • Tables 24 (Mendota Pool) and 25 (Delta-Mendota Canal): deliveries to the San Joaquin River exchange contractors and the wildlife refuges.
  • Table 26: the San Luis Canal's federal contractors (Westlands and others) and the San Felipe Division, which serves Santa Clara and San Benito.
  • Tables 27 (Tehama-Colusa Canal) and 28 (Sacramento River settlement contractors' diversions).
  Delivery lines run from the turnout to the district or refuge they serve.
• DWR Bulletin 132 (Appendix B, edition 132-24) and DWR's Historical SWP Table A Allocations: each State Water Project contractor's maximum Table A amount, every allocation announcement, and the water actually delivered each calendar year. Deliveries are actual through 2023, and DWR's projections are not used. Shown in the “Allocation vs. delivered” panel (SWP).
• DWR water-year types (CDEC report WSIHIST, “Official Year Classifications based on May 1 Runoff Forecasts”): the Sacramento Valley and San Joaquin Valley classes (wet, above normal, below normal, dry, critical) and their indices. Shown in the time bar under the map, as a strip above each feature chart (Sacramento Valley), and as a mark under each year of the System accounting chart. The report's reconstructed table, from observed runoff, differs in a few years; the official one is used. The water year in progress has no class until DWR adds it in the autumn.
• Geometry: OpenStreetMap waterways (© OpenStreetMap contributors, ODbL) and USGS NHD, with connectors drawn by hand for tunnels and pipelines. The line geometry file (data/network.geojson) is offered under the Open Database License 1.0, because it is built from OpenStreetMap data. The numbers, ledgers and notes on this map are not OpenStreetMap data.
• Reference layers: groundwater basins (Bulletin 118), water districts and groundwater sustainability agencies from the California Department of Water Resources (CC BY); county boundaries from the US Census Bureau's 2024 cartographic boundary files (public domain). Basemaps: “Gray canvas” and “Satellite” are map services and data available from U.S. Geological Survey, National Geospatial Program (public domain); “Streets” and “Gray streets” are OpenFreeMap (© OpenMapTiles, data from OpenStreetMap); “Terrain” is OpenTopoMap (CC BY-SA).
• COEQWAL (University of California; related, not a source here): CalSim 3 scenarios of the California water system under alternative rules and climates. Model results, not measurements: nothing from them is used on this map.

Terms are defined in the Glossary (header), each with the document that defines it.

Copies of cited documents are kept for citation; owners may ask for removal: williamadamsgray@gmail.com

Method

• Daily values are screened for impossible negatives and isolated one- or two-day spikes. Computed reservoir inflows and Delta outflow may be negative and are kept.
• Weekly (Mon–Sun) and monthly values are means of the daily record, published only when at least 5 of 7 or 20 days are present. A water year runs October–September and needs 330 days, or all 12 months. A turnout's delivery tables list only the months a user took water, so its water year counts the months listed, and a month the table covers with no row counts as no delivery. The monthly chart leaves that month blank.
• Volume: 1 cfs flowing for one day = 1.983 acre-feet. Storage is the last value in the period.
• River reaches show the gauge at (or nearest) their downstream end. Reaches right below a dam show the release.
• For the SWP, the Allocation vs. delivered panel compares a published plan with published deliveries, per calendar year (the SWP contract year). Allocated = maximum Table A × the year's final allocation. Delivered includes non-project water, Article 21 surplus water and carryover, so it can exceed the allocation. These are DWR's accounting figures, not flow measurements, and they never enter the measured flows on the map.
• The System accounting panel balances each reach between two measured points, per water year or calendar year. It covers the State Water Project's aqueduct reaches and the Delta-Mendota, Tehama-Colusa, Friant-Kern and Madera canals (switch with SWP / CVP / Friant). It also covers the Sacramento River between its gauges from Keswick to Wilkins Slough, where the settlement contractors divert (Sacramento R.). On the river, gauged creeks count as inflow. Many creeks have no gauge at their mouth, so a positive river balance is mostly ungauged inflow, not a missing table entry. Each balance compares the flow in at the head, plus listed inflows, less deliveries, branch outflows, and the flow out at the end, with the change in storage. The ledger reads: flow measured at the head + inflows - deliveries - outflows - storage change = "Expected" at the end of the reach. Next comes the flow measured there. The last line, "Unmeasured gain" or "Unmeasured loss", is measured minus expected. It is signed as DWR signs its own computed losses: losses negative, gains positive. A loss is seepage, evaporation, spills, unmeasured use, and measurement error together. A gain is an inflow missing from the tables (groundwater banks and wells pumping into the aqueduct). It is arithmetic on published measurements only. No number is modelled; a month that needed any approximation is marked estimated. For the State Water Project, the monthly balances to December 2020 come from DWR's monthly operations reports (2016–2017 recovered from the Wayback Machine; 2018–2020 supplied by DWR on request and marked preliminary). Those reports do not cover every reach in 2016–2017. Calendar years 2021–2023 come from CDEC daily flows and Bulletin 132 deliveries by aqueduct reach. A month built from a daily record with a day missing counts the missing days at the month's mean flow, so it is marked estimated. The Reclamation canals use Reclamation's monthly delivery tables (2005 on) and the flows on the map. None of the Reclamation canals' ends is gauged, so their unmeasured loss also holds flood and wasteway releases to rivers. Where DWR publishes its own computed losses for the same stretch (the San Luis Canal, Table 20), they are shown next to ours as a check. The two agree within 1 TAF in most months.
• Unit conversions. Several CDEC pumping-plant sensors labelled cfs actually carry acre-feet per day for some or all of their record: Devil Canyon, Gianelli, South Bay, Las Perillas, Greenspot and Crafton Hills from 2021; Oso January 2021 – December 2022; Edmonston January 2021 – December 2025 (cfs from January 2026); Buena Vista from January 2021. Each conversion was tested against an independent measurement (USGS gauges, plant ratings, reservoir balances, DWR's own reports, Bulletin 132's plant totals). DWR's own monthly volumes (supplied on request) confirmed Edmonston and Buena Vista as acre-feet per day for 2021–2024, so the Check 21 to Edmonston balance for 2023 counts as measured. The units of Oso from 2023 are still uncertain, because two independent checks disagree; the East Branch 2023 balance that uses it is flagged suspect (see Data notes). From 2023 the map leaves Oso as CDEC publishes it, until DWR states the units. The East Branch head (Edmonston less Oso) reads negative in 13 months between January 2023 and May 2026, and those months stay as published.
• Where DWR's monthly reports contradict themselves (a month's flow printed in the wrong units, or a value repeated from another month or table), the value that closes the report's own balance is used. Every such correction is logged with its evidence in the Data notes.
• The Allocation vs. delivered panel also covers the Central Valley Project (switch SWP / CVP). Allocated is each contractor's contract amount (Reclamation's “CVP Water Contractors” list, revised 30 March 2016) times the contract year's announced allocation, by Reclamation's own rule. The contract year runs March–February, and allocations are by class, from Reclamation's allocation summary. This is compared with what Reclamation's monthly delivery tables list for that contractor. Delivered counts all listed deliveries, including water-right base supply, Class 2 and Section 215 water, transfers, and non-project water, so it can exceed the allocation. Contractors the tables do not list show the allocation only, and the share of the contract amount with a record is stated. These are published plan and accounting figures, not flow measurements. The layer menu has matching map layers, CVP contractors: delivered vs allocated and SWP contractors: delivered vs allocated. They color each contractor's district for the selected year: teal where the delivered water was below the allocation, magenta where above, dashed where there is no comparison. Click a district to open its contractor in the Allocation vs. delivered panel. The CVP layer shows only districts of contractors with a delivery record. The SWP layer shows only the ten Central Valley contractors whose districts are mapped. Kern County Water Agency's agricultural contract covers eleven member districts, each showing the agency's total. A contract that covers several districts shows its total in each.
• The Pick a stretch of waterway section balances any stretch of river or canal you pick with two pins, for the period set on the time bar. It uses only the records already on the map: the flow at each end, the deliveries and diversions listed along it, mapped tributaries and branch canals, and what is left. The flow at an end is measured where a gauge sits within 250 m. Otherwise it is worked out from the gauges either side, with the loss between them spread evenly along that reach, or carried from one gauge by the listed flows. Deliveries and diversions come from Reclamation and DWR monthly tables, so they apply only at monthly steps and up. The left-over is counted only where a gauge lies on each side and the tables list every user. If negative, it is a loss or a user not listed. If positive, more water is coming in than any table lists (ungauged creeks, return flows, groundwater, pump-ins). The evaporation, rain, and seepage ranges beside it are the project's own estimates for the water year (see Estimated losses). They are in proportion to the stretch's length on each reach, and they are not added to the measured numbers. Rivers get no loss estimate, and storage change in reservoirs along the way is not included. The Monthly table button lists the stretch month by month in a window, with a total row that equals the result. Where a gauge has ended and a stretch includes an inflow with no record (Colusa Basin Drain, Sacramento Slough), its water stays in the unmeasured gain / loss, the result is flagged estimated, and dashed Estimate rows give each range for scale.
• The Measurement completeness panel scores each system for the selected water year. It gives the share of its drawn rivers and canals, by length, with a daily record, a monthly record only, an estimated record, no record that year, or no public measurement at all. It also counts how many of its turnouts have a delivery record, and how many System accounting reaches close. Length is used, not volume, because a flow nobody publishes has no volume to weigh. It is read off the records on this map; nothing in it is modelled.
• The Surface water by subbasin / GSA section lists the CVP and SWP water delivered to the contractors whose districts lie in a groundwater subbasin or GSA, by supply line, from the same Reclamation and DWR delivery records. A contractor wholly inside is summed as published. One that straddles the boundary is listed with its whole delivery and its share of area. The area-share part is marked as an estimate, because an area share is not a water share. CVP lines are water years and the SWP line calendar years, and they are never added together. Local water rights and other local supplies are not included. The Solano Project (Putah South Canal) and the Orland Project are Reclamation projects but not CVP, and are listed apart.
• A report card per SWP and CVP contractor (search, or "card" in the Allocation panel) gathers its contract, allocations, deliveries (CVP by month as Reclamation publishes them, SWP by calendar year), and record since 2001. Year in review (header) summarises a water year beside the year before and the median of the other years, each line with its own accounting year.
• Pins. A search result that is a place (a river, canal, turnout, measuring point, reservoir, the Delta, a water district, GSA or subbasin, or the districts of a contractor) has a pin button. Ctrl/Cmd+Enter pins the highlighted result. Picking a district, GSA, or subbasin (or a contractor) outlines it, with its name, and goes to it until your next click elsewhere. Typing “swp” or “cvp” finds the whole system. A pinned place stays drawn in its own color with its name until you remove it. It stays through Esc, clicks on the map, and changes of period, and it is kept in the link. The glossary’s “Show on map” pins the places a term names. A pin is only a highlight: it changes no number or panel.
• Downloads. Every chart, ledger, canal delivery list, stretch budget, and panel table downloads as CSV. Each file opens with lines naming its sources, the release, the dates covered, and the view it came from. A blank cell is no record, not zero. The loss estimates download as their own file, marked ESTIMATE, and the list of timeline events as another (see “Events on the timelines”). Public API (CC BY 4.0): api/index.html.
• The Self-reported to the State Water Board card compares nothing by arithmetic with a measured row. It shows two filed figures, with the measured volume as a separate line. The filed figures are the largest single right (the least the filings claim) and all matched rights summed (which may count the same water more than once). Report years to 2020 are calendar years, 2021 covers January–September only, and 2022 on are water years (SB 155). A water year is shown only when all 12 months are covered. A filing that enters one amount as direct diversion, as storage and as use is counted once. Water year 2025, the first CalWATRS-filed year, is provisional. Rivers the map draws without the diversion (Kings, Kern, Merced, Yuba, Mokelumne, Delta districts) list the rights with no total and no comparison.
• The D-1641 compliance indicators section compares the Delta records, month by month, with the flow requirements of Revised D-1641 as written. These are the Table 3 minimums, Table 4 X2 day counts, the export / inflow limit, and Vernalis and Rio Vista flows. Each month gets a status: met, below the minimum / above the limit as written, not determinable (an input the rule needs is not available), conditional, or not in our data. A shortfall shown there may have been excused by a temporary urgency change order or an approved variation. Those are shown as event markers, not as changes to the standard. Net Delta outflow (NDOI) is a daily water balance calculated by DWR, not a measured flow. Our values are Dayflow's. The index DWR and Reclamation computed for compliance on a given day can differ. Running averages follow the decision's footnote 2: a failed 7-day average counts all its days. The export / inflow ratio uses the 14-day inflow average by default, with a 3-day toggle. Public data cannot show which one the projects used, so the basis is always printed. The Vernalis high / low value rule is our reading of footnote 13 and reproduces the State Water Board's 2025 numbers. Gate closures at the Delta Cross Channel are not in our data. The salinity rows are measured as electrical conductivity (EC) at the stations the projects report, except where a nearby station stands in (labelled). Chloride is published only from 2006 at Banks Pumping Plant (the stand-in for C-9), 2008 at Pumping Plant #1 and 2011 at Tracy. It is never converted from EC. EC is in mmhos/cm as D-1641 prints it (1 mmhos/cm = 1 dS/m = 1,000 uS/cm). Running means are compared at 0.01 mmhos/cm, as the projects report them. A day that only that rounding saves is tagged within rounding. Where D-1641 and the 2018 Bay-Delta Plan differ for the southern Delta, D-1641 is the test and the plan's 1.0 mmhos/cm (written 1.0 dS/m) is a dashed reference line.
• Values marked estimated come from a fallback expression or are missing a minor term. Example: Banks pumping stands in for the Bethany-to-O'Neill reach before South Bay pumping appears on CDEC in 2021.

Estimated losses (modelled, not measured)

The System accounting panel splits each reach's measured “unmeasured loss” into what evaporation, seepage, and rain on the water could explain. These are estimates made by this project, not measurements and not figures published by DWR. They are shown as ranges with a confidence grade, beside the measured balance, and never change it. Ranges cover the stated assumptions only; the true value can lie outside them. Not for operational, legal, or financial use.

• Evaporation (grade A: computed from measured inputs). Daily reference evapotranspiration from gridMET (University of Idaho, ~4 km grid; Abatzoglou, 2013, International Journal of Climatology 33: 121–131; public domain, CC0) is taken under the open stretches of each reach. It is multiplied by the open-water coefficient from FAO Irrigation and Drainage Paper 56, Table 12 (0.65–1.25, central 1.05), and by the water surface. Canal width is measured from OpenStreetMap's mapped canal water (traced from aerial imagery): about 54 m on the San Luis Canal, 40 m on the main aqueduct, 25 m on the East Branch. Pipelines and tunnels are left out. Lakes inside a reach (Bethany Reservoir, Silverwood Lake, Lake Perris) use their mapped shoreline, about full pool, so their evaporation is high when a lake is drawn down. As a check, for Silverwood and Perris together, DWR's own computed losses are in its monthly reports, which also net in rain and natural inflow. Over water years 2001–2015 they are a median 1.08 of our central net estimate (evaporation net of rain), and fall inside our range in 6 of those 15 years. The other years scatter widely, as natural inflow makes Silverwood a net gain in DWR's figures in several wet years.
• Rain on the water surface (a gain; grade A: computed from measured inputs). Daily precipitation from the same gridMET cells (University of Idaho, ~4 km grid) is multiplied by the same water surface as the evaporation. There is no coefficient, because rain falling on water is a direct input. Estimates carry the balance's signs: evaporation and seepage are negative, the rain positive. The estimate shown is their net sum. Only rain on the water itself counts; runoff from the canal banks and the land beside it is not estimated. Lakes use their mapped shoreline, about full pool, so their rain is high when a lake is drawn down. The canal's range follows the same open and unknown lengths as evaporation, so the largest area gives the most rain: the top of its range, and the bottom of the evaporation range.
• Seepage (grade C: rate of a similar canal). No seepage test of the California Aqueduct is published. The rate is the Central Arizona Project's, a concrete-lined aqueduct: about 9,000 acre-feet a year on 336 miles, per CAP. It is scaled per mile of open canal for the smaller loss and per area of wetted lining for the larger loss. No central value is given.
• Measurement uncertainty. The residual is a small difference between large measured flows, so a small meter error moves it a lot. The panel shows how far it could move if every measured term were off by 2% or 5%. These accuracies are assumptions. DWR does not publish the accuracy of its plant and turnout meters.
• Reading. If what is left after evaporation, seepage, and rain is within the 2% uncertainty, the reach is “consistent with these losses plus meter error”. If it is clearly positive (a gain), an inflow is missing from the published tables (groundwater pumped into the aqueduct, natural inflow to a lake). If it is clearly negative (a loss), something else left the reach: unmeasured diversions, spills, larger meter errors, or an outflow the tables do not list.
• Reclamation canals (Delta-Mendota, Friant-Kern, Madera). Seepage uses Reclamation's own tests, from Linings for Irrigation Canals (1963). Ponding on two reaches of the Friant-Kern Canal in 1950 gave about 0.07 ft of water a day, through both its concrete and its compacted-earth lining. Reinforced concrete on the Contra Costa Canal gave 0.008 and 0.09 ft/day (1958). The Delta-Mendota Canal's earth lining gave 0.009 ft/day (permeameter). The large case for earth lining is the 0.40–0.43 ft/day of unlined control reaches in the same report. Each canal's split between concrete and earth lining is Reclamation's (Friant-Kern 85% concrete, Madera 79% earth, Delta-Mendota concrete to milepost 98.64 of 116.6). All these tests are of linings a few years old, so today's seepage may be higher. These canals run low or dry for months; in a month without water nothing is estimated. Otherwise the wetted area is scaled by the flow against the canal's design flow: in proportion in the small case, by the square root in the central case (an assumption), and the full section in the large case. OpenStreetMap does not map the Delta-Mendota Canal's water, so its width is from a published section of its earth reach (grade B). The Tehama-Colusa Canal has neither a mapped water surface nor a published section, so it is not estimated.
• Not estimated: the Coastal Branch (its water surface is not mapped), runoff into the canals, spills, and unmeasured diversions.

How to cite

Usage counts

This site counts page views and which features are opened, using GoatCounter. It sets no cookies, stores no IP addresses, and collects no personal data; what you type in a search box is never sent. The counts are totals only. GoatCounter's privacy policy.

The site also keeps its own daily totals of the same page views and feature names, on a small Google Cloud service run for this map. It keeps a count per day and per name, nothing else: no cookies, no IP addresses, no browser details (only whether a page view came from a known crawler) and nothing you type. Frozen release copies and pages shown inside another site are not counted. To leave a browser out of these counts and GoatCounter's, open usage/off once in it (a flag in the browser's storage holding only the word "off"; usage/on undoes it).

Known gaps

• SWP aqueduct plants south of the Delta are on CDEC only from 2021 (Dos Amigos from 2010). Earlier years come from DWR's monthly operations reports, which run to December 2020: DWR posts them to 2017 and supplied 2018–2020 on request, marked preliminary (see Data notes). A few 2016–2020 tables are not read or are missing a page (see Data notes). Months missing from both sources show as no record.
• The lower Delta-Mendota Canal and the Glenn-Colusa Canal have no public daily record.
• The Hetch Hetchy Aqueduct shows CDEC's Hetch Hetchy "diversion" series, which starts in September 2021. CDEC does not say whether it is the whole aqueduct flow. The Mokelumne Aqueduct (EBMUD) and Los Vaqueros Reservoir have no public daily record.
• D-1641 indicators:
  • No Delta Cross Channel gate log.
  • No Chipps Island EC record. Mallard Island stands in from January 2008, with nothing before that.
  • No Rio Vista or X2 after Dayflow's last water year.
  • No Bulletin 120 for June or the January issue.
  • The pulse-flow dates are the default 15 April – 15 May window.
  • Salinity: no chloride before May 2006 (Banks Pumping Plant, standing in for C-9) or April 2008 at Pumping Plant #1, and none at Antioch, Barker Slough or Cache Slough.
  • Salinity: no accepted Brandt Bridge EC for 1 March – 19 May 2005 or Old River at Tracy Road EC for 22 January 2004 – 4 April 2005. No Prisoners Point EC for 1 April – 18 May 2006.
  • Salinity: no published Suisun Marsh high-tide EC, and no Suisun gate certifications.
• Delta channels: on the Delivery Map, Georgiana Slough appears in the Delta panel but is not drawn; the Tributaries Map's "Delta channel flows" layer draws it with 34 other measured channels. Twenty USGS Delta channel records, the Cross Channel and Georgiana Slough among them, end on 27-30 September 2025, when USGS handed the stations to Reclamation, whose flows are provisional (see Data notes). Old River, Middle River and 13 other channels continue, with USGS approval about two years behind.
• Diversions along a canal are not subtracted. A canal shows the flow at its measuring point, usually its head.

Central Valley Surfacewater Map

List of features

Historical surface water in the Central Valley (the State Water Project, the Central Valley Project, the Friant Division, rivers, reservoirs, and the Delta), from published measurements, daily back to water year 2001.

1. Search and pins. Find a canal, reservoir, district, GSA, subbasin, or contractor by name; pin several to keep them highlighted together. link to feature |
2. Allocation vs. delivered. What each SWP and CVP contractor was allocated and what it received. link to feature |
3. Wild & Scenic rivers. Every federal and state Wild and Scenic river in California as a map layer, with a card per river: the laws that designated it, the plans and court decisions that bear on it, and how it relates to the rivers drawn here. A designation is a legal status, not a measurement, and sets no flow requirement. link to feature
4. System accounting. Reach-by-reach ledgers of the canals and the Sacramento River: water in, deliveries out, and the unmeasured gain or loss, beside DWR's own figures and labelled estimates of losses. link to feature |
5. Pick a stretch of waterway. Drop two pins on any river or canal to get the budget between them. link to feature |
6. Delta water balance. Inflow, exports, and outflow. link to feature |
7. D-1641 compliance indicators. The Bay-Delta outflow, X2, export, and Vernalis requirements of Revised D-1641 and its salinity objectives (chloride, agricultural EC, Suisun Marsh). Each is checked month by month against the published records, with a daily chart. link to feature |
8. Surface water by subbasin / GSA. Project deliveries placed on groundwater subbasins and GSAs, with what the GSAs reported. link to feature |
9. Report cards and downloads. One page per contractor. Every chart and table downloads as CSV. “Cite” gives a frozen release to quote. link to feature |

Measured data only. Estimates are labelled as estimates and never added into measured totals. Where nothing is published, the map says “not publicly measured”.

Year in review

Data notes

Every error found in a published source, and every bug found in this project's own processing, that changed (or could have changed) a number on this map. Each note says what it looked like, how it was settled against independent numbers, and what the map does now. Gaps in what the agencies publish that bias a balance are listed too.

Corrected: the map uses a value confirmed by another published table or gauge. Left out: the bad values are dropped and show as no record. Flagged: kept, with the doubt noted where it shows. Open: not settled yet; nothing has been filled in with a guess.

Glossary

The terms this map uses, in plain language, each paraphrased from the document that defines it. Where the map uses a term more narrowly than the document does, that is said under “On this map”. Underlined terms in the panels open their entry.