Zarafshan Valley Groundwater Model
→ ↓ Space
Next slide
Следующий слайд
← ↑
Previous slide
Предыдущий слайд
Home / End
First / last slide
Первый / последний слайд
C
Contents rail — chapters and sections
Оглавление — главы и разделы
O
Overview grid — click a slide to jump
Обзор слайдов — щёлкните, чтобы перейти
N
Presenter notes
Заметки докладчика
P
Print / export to PDF
Печать / экспорт в PDF
L
Switch language EN / RU
Переключить язык EN / RU
?
This help
Эта справка
Ministry of Mining Industry and Geology of the Republic of Uzbekistan
Regional groundwater model of the Zarafshan valley · MODFLOW 6
Modelling and analyticshydrosolutions GmbH

Status review · September 2026

Groundwater in the Zarafshan valley

A regional groundwater model of the valley between the Ravathoji headworks and the Khazara gauge: the water balance it computes, what the balance measured in 1968 says is wrong with it, and the seven steps that follow.

Overview

Seven parts

  • 1 Introduction: the valley and the pressures on its groundwater
  • 2 Data and methods: what the model is built from
  • 3 Results: the water balance of the valley, of the groundwater, and month by month
  • 4 What is wrong, and what the 1968 balance says
  • 5 The plan: seven steps, one change and one test each
  • 6 Conclusions
  • 7 Appendix: the coupled system, the scenarios, questions to the experts, method details

Keys: → next slide, C contents, O overview, N notes with sources, L language, P print to PDF.

Summary

Key messages

Almost all groundwater in this valley is water that leaked out of the river, the canals and the fields. Rain and the mountains add little. That was measured in 1968 and it is still true.

The model computes a groundwater balance well under half the size of the one measured in 1968, and returns to the river less than a tenth of what came back to the surface then. Every other problem in this review follows from that.

The model itself is sound: it runs in seconds, reproduces the shape of the water table and the seasonal rhythm, and is handed over and reproducible. What it lacks is the right water in the right places.

The plan puts it there in seven steps, each with a map, one change and one test. The first four need no new data. After that the model is ready for the scenarios agreed with MMIG.

1 Introduction

A valley that lives on one body of groundwater

From the headworks (1) to the last gauge (2) the river runs through the most intensively irrigated land of Uzbekistan. Under the fields lies one shallow body of groundwater in the river gravels that the canals, the fields, the river and the towns all draw on.

Samarkand, Kattakurgan and Navoi (3, 4, 5) take their drinking water from it. Farmers pump it where canals do not reach. The river gains from it in winter and loses to it in summer.

Permits, monitoring and new well fields are decided district by district, but the groundwater does not know district boundaries. A regional model is the only way to see the whole system at once.

A valley that lives on one body of groundwater

Source: Positions: MoW station file (data/mow/mow_stations_swb.gpkg), WorldPop city cells, OSM; fan circle and northern belt drawn from the model grid (schematic). Water bodies: zarafshan-water-body-delineation register.

1 Introduction

A dense web of canals feeds and drains the valley

A dense web of canals feeds and drains the valley

Finding. From the Ravathoji hydroworks the river is split into the Dargom, Eski Anhor and Tuyatartar canals and the Zarafshan's own branches; further down the Narpay, Bulungur, Yangiariq and Kattakurgan canals and hundreds of smaller ones spread the water across the plain, and a second web of drains and ditches carries the surplus back towards the river.

What it means. Every canal leaks into the aquifer and every drain skims it off again. Groundwater in this valley is a by-product of this network, which is why the model cannot be understood without it.

Source: OpenStreetMap waterways (Uzbekistan extract, September 2026; incomplete in places), SRTM relief, HydroLAKES reservoirs, model outline.

1 Introduction

The study area: 230 km of valley and everything the model must account for

The study area: 230 km of valley and everything the model must account for

Finding. The model covers the valley floor and terraces, 10 208 km², from the Ravathoji hydroworks to the Khazara gauge. The Zarafshan splits into the Oqdaryo and Qaradaryo branches, feeds the canal systems and the Kattakurgan reservoir, and receives three ungauged tributaries from the north, whose basins lie outside the model.

What it means. Everything the model says is regional: districts and river reaches, not individual wells. The mountain catchments enter only as inflows at the model edge.

Source: OpenStreetMap tiles and canals; SWB outline and grid; tributary basins (hydra-shed); HydroLAKES; inflow points; WorldPop city cells.

1 Introduction

Four pressures on the same aquifer

Population growth

A third more people in the model area since 2010. Domestic pumping grows with them, and the towns' well fields sit in the most permeable part of the aquifer.

Climate variability and change

Rainfall over the valley varied between 178 and 471 mm a year in 2010–2024. Warmer summers raise crop demand; the snow-fed river peak is expected to shift earlier.

Unmeasured abstraction

The register lists eight times more drinking and industrial pumping than population norms explain, without well coordinates. Agricultural and unregistered pumping are not measured at all.

River diversion

Two thirds of the river is diverted at Ravathoji before it enters the valley. What is left in the channel sets how much the aquifer can exchange with it.

1 Introduction

Where the water table is falling

Where the water table is falling

Finding. Of 102 monitoring wells with at least eight years of data, 38 show levels falling by more than 0.1 m a year and only 5 rising. The strongest declines cluster south of Samarkand and in the eastern terraces.

What it means. The median trend is close to zero, so this is not a valley-wide depletion yet. It is a set of local declines that a regional model must be able to attribute to pumping, canal changes or dry years.

Source: data/observations/regional_gw_observations_long.csv, quality-checked values, straight-line trend per well (measured here).

1 Introduction

How much is really pumped? Nobody knows precisely

For the thirteen districts around Samarkand the MMIG register reports 27.5 m³/s of drinking and industrial pumping. The number of people living there explains 3.6 m³/s. The gap sits almost entirely in the three districts with the large well fields (1).

Those well fields send water by pipe to Navoi and Bukhara (2, 3), and part of the city's water comes back through the sewer (4). Without well positions, yearly volumes and the piped amounts, the model cannot say where this pumping takes water from. Farm pumping is in neither number.

How much is really pumped? Nobody knows precisely

Source: Positions: WorldPop city cell, MoW station file, model grid (fan circle schematic). Damkhodja and Siyob: Juraev et al. 2021 (docs/soviet_literature_crosswalk_2026-09-15.md §6.4). Well field 5: coordinates given by the project owner on 2026-09-16. Arrow paths are schematic.

1 Introduction

What the model has to answer

  • How much water reaches the aquifer from canals, fields and rain, and how much leaves through pumping, evaporation and the river?
  • Where and when does the river gain groundwater, and where does it lose it?
  • What happens to levels if pumping grows, if a dry decade comes, or if canal losses are reduced?
  • Which of these answers can be trusted today, and which need more data?

The next three parts follow these questions: the data and methods, the results, and the limits.

2 Data and methods

Everything the model is built from

Everything the model is built from

Finding. Eleven datasets from MMIG, UzHydromet, the Ministry of Water Resources and global satellite products. Groundwater levels cover 2000–2024, river flow 2010–2024, canal records 2014–2022 with reliable flow below the intakes only for 2014–2017.

What it means. Every input is traceable to one of these bars. The short bars are the weak points: the canal record and the single-year population snapshot.

Source: data/README.md and the per-folder READMEs; coverage as recorded there.

2 Data and methods

How water moves through the valley aquifer

How water moves through the valley aquifer

Finding. One open water table between the land surface and the bedrock. It is fed by canal and field losses, a little rain, the fans of the northern tributaries and groundwater arriving from upstream; it is drained by the river, evaporation from shallow groundwater and crops, and pumping.

What it means. This concept is what the model implements. Two of its elements, the bedrock depth and the foothill margins, are also where the model is weakest, as chapter 4 shows.

Source: Conceptual drawing after the MMIG cross-sections and the model concept; no model result is shown.

2 Data and methods

Five geological zones carry the aquifer properties

Five geological zones carry the aquifer properties

Finding. The geological map of the young deposits was traced and transferred to the model grid: recent river deposits along the Zarafshan, two generations of older terrace deposits on both flanks, small patches of the oldest deposits, and bedrock, which is left out.

What it means. Wells drawing from the young deposits and from the older rock beneath them show no systematic difference in water level, so one layer is a sound choice. The zones follow MMIG's field mapping and can be refined with MMIG geologists.

Source: data/model/aquifer_properties.npz zone_grid (measured square counts).

2 Data and methods

How fast water moves through the ground, and how much the ground stores

How fast water moves through the ground, and how much the ground stores

Finding. Left: the permeability of each geological zone, that is how many metres a day water can move through it, from pumping tests in the MMIG well registry: about 30 in the recent river gravels, 5 in the younger terraces, 2 in the older terraces, under 1 in the oldest deposits. Right: the share of the ground that drains as water when the water table falls, from published values: a quarter for the river gravels, down to a few percent for the oldest deposits.

What it means. Both are starting values, one per zone. The Soviet pumping tests put the fan gravels at 100 to 200 m a day, three to six times higher (chapter 4), and the month-by-month tests show the storage values are too high (appendix D).

Source: data/model/aquifer_properties.npz (K by zone, m/d) and aquifer_zone_lookup.json (Sy by zone). Both are starting values: the Soviet pumping tests give 100–200 m/d for the fan gravels (chapter 4) and the month-by-month tests show the storage values are too high (appendix D).

2 Data and methods

Where groundwater levels are measured

Where groundwater levels are measured

Finding. 398 wells passed quality control: 148 with monthly time series and 250 with a single measurement. Coverage is dense around Samarkand and along the northern edge, sparse west of Kattakurgan.

What it means. These wells are the yardstick for the model. Where they are sparse the model is weakly checked; where several sit in the same 1 km square they cannot all be matched.

Source: data/observations/calibration_targets_steady_state.csv (MMIG registry, ZRB monitoring wells, UGV time series).

2 Data and methods

The river system: where water is taken out and where it comes back

The river system: where water is taken out and where it comes back

Finding. At the headworks (1) the river is split; two large canals leave the valley for other regions. The rest parts into two branches around the Miankal island (2), feeds the reservoir and the Narpay canal at Damkhodja (3), drops to 5 m³/s below the Narpay works (4), rejoins at Khatyrchi (5), carries 41 m³/s again at Ziatdin (6), feeds two more canals at Karmana (7), and leaves the valley at Khazara (9) with 23 m³/s.

What it means. The model follows only the dark blue line and, until now, took no water out of it and put none back below the headworks. These nine points are where the balance is checked.

Source: Positions: MoW station file; flows quoted are 2014–17 means of the MoW decadal record (GAPS.md §5 Iteration 6, read). Canal geometry: delineation register.

2 Data and methods

Most of the river is diverted before it enters the valley

Most of the river is diverted before it enters the valley

Finding. The Zarafshan brings 134 m³/s to Ravathoji on average. Transfers to the Jizzakh and Kashkadarya basins take 39 m³/s, the valley canals another 49 m³/s. The river below the intakes carries 46 m³/s, a third of the inflow.

What it means. The model's river starts with that lowest line. Canal water enters the aquifer as recharge through the fields, which is why the loss fractions matter so much.

Source: data/mow/mainstem_sfr_inflow.npz (UzHydromet daily discharge, Ministry of Water canal records 2014–2017, other years by monthly ratios).

2 Data and methods

Where the water is used

Where the water is used

Finding. Landsat maps for 2016–2020 give the irrigated share of every 1 km square: 3 272 km² inside the model area, concentrated on the valley floor and thinning towards the terraces.

What it means. This share splits recharge and water loss to the air between irrigated and rain-fed land. It is the most decisive single input of the water balance.

Source: data/irrigation/regional_irrigation_fraction.npz; area = sum of the irrigated shares over active squares.

2 Data and methods

What goes in and what is drawn out, every month

What goes in and what is drawn out, every month

Finding. Recharge combines rain seeping into the ground with canal and field losses, weighted by the irrigated share. Water loss to the air combines crop use, evaporation from shallow groundwater and from open reservoirs. Both maps exist for each of the 180 months.

What it means. The share of canal and field losses that reaches the water table, set at 45 percent, is the key assumption. Chapter 3 shows what the 1968 balance says about it.

Source: data/model/regional_rcha_evt_arrays.npz, long-term average maps, converted from m/d to mm/yr.

2 Data and methods

Where groundwater is pumped in the model

Where groundwater is pumped in the model

Finding. Domestic and industrial pumping is spread with 100 m population data and Uzbek per-capita norms, 3.4 m³/s in total for the April model. The September test version instead uses the full register value of 37.8 m³/s, smoothed within districts.

What it means. Neither placement is right: the first is too small, the second puts well-field pumping where the population is rather than where the wells are.

Source: data/population/pumping_mean_annual.npz and population_regional_2024.npz (WorldPop 2020, projected).

2 Data and methods

How the model talks to its surroundings

How the model talks to its surroundings

Finding. The river is computed along 300 channel sections, so its water level responds to inflow and to exchange with the aquifer. Fixed water levels close the two valley ends, 75 reservoir squares exchange water with the aquifer depending on their level, and pumping is applied square by square.

What it means. The northern and southern foothill edges are closed to flow and receive almost no recharge. That choice is revisited in chapter 4.

Source: data/model/regional_model_setup_v2.json (river and fixed-level data), regional_reservoir_mask.npy, regional_wel_data.npz.

2 Data and methods

The model in numbers

72 × 223 squares of 1 kmOne layer, 10 208 active squares
180 monthly periodsMonth-by-month simulation 2010–2024
300 river reachesRiver computed along its channel, 40–500 m wide
381 wellsObservation wells with a long-term mean level
17 849 monthly observationsMonthly level observations, 144 wells
< 1 s per long-term runEnables thousands of test runs

Built with MODFLOW 6, the standard open-source groundwater code, and open-source tools throughout.

3 Results: the water balance

The valley water balance, and what is left for the aquifer

The valley water balance, and what is left for the aquifer

Finding. Inflows of 8.2 km³ a year against 6.8 km³ of measured or estimated outflows leave 1.4 km³ unaccounted. With the assumed loss fractions, 0.96 km³ of that recharges the aquifer and 0.46 km³ is lost in the soil zone. The Uzhydromet record at Navoi (2010–2020 mean 0.80 km³/yr) sits just above the 0.72 km³/yr booked at Khazara from the four gauged years.

What it means. The aquifer's recharge is the one term no gauge measures. It is derived as a remainder, which is why every other term must be as good as the data allow.

Source: data/mow/regional_water_balance_constraints.json. Recharge = 45 % of canal and field losses plus 10 % of precipitation. Navoi check: Uzhydromet monthly discharges 1967–2020 (data/uzhydromet/zeravshan_downstream_monthly_q_m3s.csv), May 2016 outlier corrected.

3 Results: the water balance

Three balances, one valley: how much water reaches the ground, and where it goes

Three balances, one valley: how much water reaches the ground, and where it goes

Finding. Left: of the roughly 260 m³/s that enter the model area, most evaporates from fields and rain-fed land or leaves in canals to other regions; 45 m³/s cannot be accounted for, and about 30 of that is the water that seeps down to the groundwater. Middle: the model's groundwater receives 29 m³/s and gives back 24 by evaporation, 4.5 to the river, 3.4 to pumps and 1.4 underground to the west; it also loses 3.5 m³/s of stored water every year. Right: in 1968 the groundwater received 82 m³/s, almost all leaked from the river, canals and fields, and gave 62 of it back to the surface as springs, drains and river flow.

What it means. The model's groundwater moves well under half the water it moved in 1968, and returns to the surface less than a tenth of what returned then. Either the valley has changed that much, or the model puts water in through the wrong doors and lets too little out through the right one. The following slides show it is mostly the second.

Source: A: data/mow/regional_water_balance_constraints.json (2014–17 reference years; km³/yr converted to m³/s). B: net package fluxes of data/model/regional_tr_workspace_p2/zrb_regional_ss.cbb via gwswi.model_v2.read_water_budget_timeseries, mean of 180 monthly periods (measured); pumping = the WEL term, which in this transient workspace carries no tributary injection (WEL inflow 0 in every period, measured); storage = STO-SY + STO-SS. C: Гидрогеология СССР т. XXXIX Table 13 p. 326 and Table 8 unit 5-а (read). Panel A includes what never reaches the aquifer; panels B and C are the aquifer alone.

3 Results: the water balance

The 1968 balance, as printed in 1971

The 1968 balance, as printed in 1971

Finding. Volume XXXIX of Hydrogeology of the USSR (Nedra, 1971; data to 1968) carries a closed groundwater balance of the middle Zarafshan valley, resource tables for eight hydrogeological units, the permeability of the Samarkand fan and the seasonal swing of the water table by unit. Read together with its Tajik companion and the Amu Darya hydrology volume of the same series, it is the reference the model never had.

What it means. Ninety-eight percent of the inflow is water lost from the river, canals and fields; three quarters of the outflow comes back to the surface as springs, drains and river gain. That is the shape the model's balance has to take.

Source: Гидрогеология СССР, т. XXXIX, Узбекская ССР. Недра, 1971, p. 326 (scan from geokniga.org, cropped). Inflow: underground inflow from the mountains 1.7, precipitation 0.4, losses of surface waters 79.4; outflow: underground outflow 0.5, evaporation 19.4, emergence into surface watercourses 61.9 m³/s.

3 Results: the water balance

The groundwater month by month: filled in early summer, emptied through autumn

The groundwater month by month: filled in early summer, emptied through autumn

Finding. Seepage from canals and fields peaks in June and July at 60 to 67 m³/s; evaporation from shallow groundwater follows in August at 57. From May to July the river loses water into the ground; in every other month the ground drains into the river. Storage fills in early summer and empties from August to October, and over a whole year the model loses 3.5 m³/s of stored water: its water table falls in every year but 2019.

What it means. The timing is right: this is the rhythm the canals impose, and it is the rhythm the wells show. The size is not: this cycle moves the water table by centimetres where the wells move by a metre and the Soviet observers saw 13 to 18 m on the fan. The steady loss of storage is the model's way of saying it receives too little water.

Source: Net package fluxes of the April transient model, data/model/regional_tr_workspace_p2/zrb_regional_ss.cbb, read with gwswi.model_v2.read_water_budget_timeseries; monthly means over 2010–2024 by calendar month, and yearly means of the storage terms STO-SY + STO-SS (measured at render). Sign convention: a positive storage term in MODFLOW is water released from storage, shown here as an inflow. The WEL term is pumping only; this workspace carries no tributary injection (measured).

3 Results: the water balance

The river gains groundwater most of the year and loses it at the irrigation peak

The river gains groundwater most of the year and loses it at the irrigation peak

Finding. Over the whole river and all fifteen years the aquifer releases about 10 m³/s to the Zarafshan in the gaining months. In June and July, when canal water raises the water table, the direction reverses and the river loses up to 30 m³/s to the aquifer.

What it means. This seasonal reversal is the headline result for management. Its magnitude is uncertain because it is constrained by levels only; the measured river flow at Khazara is the missing check.

Source: data/model/regional_tr_workspace_p2 water accounts, river exchange summed over all sections (run of 2026-09-14; measured here).

3 Results: the water balance

The simulated groundwater surface follows the valley

The simulated groundwater surface follows the valley

Finding. With average 2010–2024 inputs the model reproduces the regional gradient from about 900 m near Ravathoji to about 330 m at Khazara, with the flat central plain around Kattakurgan and steeper slopes on the flanks.

What it means. The regional picture is right. The question is how well individual wells are matched, and where not.

Source: data/model/regional_ss_workspace/zrb_regional_ss.hds (run of 2026-09-14).

3 Results: the water balance

Where the model is too high or too low

Where the model is too high or too low

Finding. Against 381 wells the untuned model is 7 m too low on average; the typical error at a well is 19 m, and the largest errors push the overall error measure to 30 m. Red dots south of Samarkand are too high; the dark blue cluster on the northern edge is 30 to 60 m too low.

What it means. The pattern is structural: the northern cluster lies outside the mapped river deposits, in the foothill belt the model treats as almost dry.

Source: model water levels at the well squares; wells from calibration_targets_steady_state.csv without the flagged inconsistent wells (measured here).

3 Results: the water balance

Fifteen years month by month

Fifteen years month by month

Finding. The month-by-month model runs all 180 months without any part drying out and with a water-accounting error below 0.02 percent. Against 15 491 observations, leaving out the first two years of settling-in, the typical error is 9.5 m. Some wells are matched closely, others are offset by several metres, and one shows the seasonal swing the others lack.

What it means. The model can already say how the regional water table responds to wet and dry years. It cannot yet reproduce the seasonal rhythm of most wells.

Source: data/model/transient_residuals.csv (run of 2026-09-14).

3 Results: the water balance

The seasonal swing is too small

At 139 wells the simulated yearly rise and fall of the water table is on average a fifth of the observed one. Only a handful of wells swing too much.

Five hundred month-by-month runs traced this to the storage settings: less releasable storage in the river deposits and terraces brings the seasonal rise and fall closer to the observations.

The seasonal swing is too small

Source: Yearly high minus low per well, averaged over 2012–2024 (measured here from the month-by-month results); 480 of 500 test runs succeeded, best seasonal error 0.23 m against 0.39 m for the starting values.

4 What is wrong, and what 1968 says

What the balances say is wrong, and where

The three balances point at five places, numbered on the map. Everywhere, the model takes water from rain and from injected inflow where the 1968 balance has leaks from the river, canals and fields, and it returns far too little to the surface.

  • 1 Samarkand well fields: the register's pumping drops the model's water table by 30 to 140 m; the wells show no drop.
  • 2 Northern foothills: 153 wells, forty percent of all, lie outside the river gravels; the model puts them 20 to 100 m too low.
  • 3 The lower river: 50 m³/s leave at Khazara where 23 are measured, and nothing feeds the 36 m³/s that return between Narpay and Ziatdin.
  • 4, 5 Southern mountains and western edge: closed in the model; the 1968 balance gives both small but definite numbers.
What the balances say is wrong, and where

Source: regional_ss_workspace heads vs calibration_targets_steady_state.csv (measured here); GAPS.md §5 Iterations 3–5; gate bands config/ss_gate_bands.json.

4 What is wrong, and what 1968 says

The same comparison, term by term

The same comparison, term by term

Finding. Four terms differ by an order of magnitude. Precipitation recharge: 0.4 m³/s in 1968 against 10.9 in the model, which applies 10 percent of rainfall everywhere. Mountain inflow: 1.7 against 21.7 injected at the northern and eastern edges. Surface-water losses: 79 against 31, because the model books river-bed losses and canal seepage only in part. Emergence: 62 against 36. Evaporation from groundwater is the one term the two agree on.

What it means. The model puts water in through the wrong doors and lets too little out through the right one. Moving the recharge from rain and edge injection to river and canal losses on the fan, and letting more of it drain back, is the correction the 1968 balance asks for, and it is what the river gauges below Narpay have been saying.

Source: Гидрогеология СССР т. XXXIX pp. 326 and 270–273 (read from the scan); docs/steady_state_water_balance_ledger.md rows 12, 20–26 and GAPS.md:1523-1524 for the candidate. Model 'mountain inflow' = northern-tributary and Ravathoji injection wells; 'pumping' in 1968 = current use of the fan segment (Table 8, unit 5-а).

4 What is wrong, and what 1968 says

Where the groundwater comes back up, and why the river gains below Narpay

Where the groundwater comes back up, and why the river gains below Narpay

Finding. The fan between Ravathoji and the Tugainy gauge takes in 20 m³/s of river and canal water and gives nothing back; the water table there lies 70 m down and swings 13 to 18 m a year. From the fan toe at Samarkand to Tashrabad the same water comes back up: 16 m³/s as the Karasu springs, 20 between Chapanata and Ishtykhan, 18 between Ishtykhan and Tashrabad, on gravels of 100 to 200 m/d. The Kermine-Kenimekh oasis below adds 5.4 m³/s of drainage to the river.

What it means. The 36 m³/s that enter the lower river unseen by our gauges were already there in 1968, and the Soviet gauges named them: the Talligulyan canal returned 32.7 m³/s into the Karadarya (1951–70 mean, 35 to 43 in winter), the Bulungur spillway 7.4 into the Akdarya, Miankol-Khatyrchi 5.5. The model needs a high-leakiness river bed on the fan, its drains and springs where the Soviet units put them, and the river allowed to gain 40 to 60 m³/s between Samarkand and Karmana.

Source: Гидрогеология СССР т. XXXIX, Table 8 pp. 270–273 (read from the scan); K is the printed filtration coefficient of the unit. Unit boundaries follow the Soviet gauge and distributor names; their modern positions are assumed.

4 What is wrong, and what 1968 says

Where the water comes back into the river

Where the water comes back into the river

Finding. Between the Narpay works (N) and Ziatdin (Z) the river grows from 5 to 41 m³/s. The Soviet gauges recorded where that water came from: the Talligulyan spill alone returned 32.7 m³/s into the Kara-Darya (1), the Bulungur and Miankal-Khatyrchi returns another 13 (2, 3). Today's drainage surveys give 17 to 19 m³/s from the Samarkand collectors (4) and 9 to 11 from the Navoi collectors (5).

What it means. The 36 m³/s the model cannot place were there in 1968 and are there today. The model has to put them back at these points.

Source: Returns: Ресурсы поверхностных вод 1976 Table 4 (1951–70 means), Shodiev & Chembarisov 2021 (docs/soviet_literature_crosswalk_2026-09-15.md §6.2, §11); river flows: MoW 2014–17 (GAPS.md §5 Iteration 6). Talligulyan course from the ZBUIS linear scheme, drawn straight (assumed); collector positions at the reach midpoints (assumed).

4 What is wrong, and what 1968 says

The one Hydromet series inside the model: Navoi, 1967 to 2020

The one Hydromet series inside the model: Navoi, 1967 to 2020

Finding. Recovered from the project archive: Uzhydromet monthly discharges at the Navoi post (1967–2020) and at the Ak-Karadarya hydrowork (1924–2020). At Navoi the river fell from 52.8 m³/s in 1967–75 to 25.3 in 2010–20, with 2010–20 annual means between 15 and 39. The Karadarya below the divider carried 35 m³/s in the same decade.

What it means. The Khazara agreed range of 15 to 28 m³/s, set from four years of Ministry of Water gauges, is confirmed by an independent 54-year record 30 km upstream. The model's carried inflow below the divider, 45.8 m³/s from 2014–17 canal records, is about 20 percent above the Hydromet decade mean and should come down. The 1968 balance and the modern state now share a measured river.

Source: Uzhydromet monthly discharges, workbook of 16 May 2023 (data/uzhydromet/zeravshan_downstream_monthly_q_m3s.csv; years with at least 11 months; Navoi May 2016 corrected from 347.0 to 34.7). Model inflow: data/mow/mainstem_sfr_inflow.npz; band: config/ss_gate_bands.json.

4 What is wrong, and what 1968 says

It is not the climate alone: a century of measured discharge

It is not the climate alone: a century of measured discharge

Finding. At the mountain exit the Zarafshan has been gauged since 1914. Its annual mean was 161 m³/s over 1915–1975 and 135 over 2010–2024: a decline of about 15 percent, statistically significant, part of it upstream use in Tajikistan. Inside the valley the same river fell from 53 m³/s at Navoi in 1967–75 to 25 in 2010–20, and at Khazara from 79 (1925–75) to 22 (2014–18): a loss of one half to two thirds.

What it means. Roughly a fifth of the downstream decline can be traced to the mountains; the rest happened between Ravathoji and Khazara, in the canals, the fields, the reservoirs and the pumps. Falling water levels in the valley are therefore first a question of allocation and abstraction, and the model has to treat them that way. Climate change is real in these series, but it is not the main story.

Source: Uzhydromet station 17461 monthly 1914–2011 (data/uzhydromet/Ravathoji_17461_monthly_1914_2011.csv; complete years only) joined with the Uzhydromet daily series 2010–2024 (Inflow_Rovatkhodzha_daily_2010_2024_ts_m3s.csv; the two overlap in 2010 at 169 vs 154 m³/s); Navoi and Karadarya from the Uzhydromet workbook of 16 May 2023; Khazara 1963–75 from Основные гидрологические характеристики (1976, 1980) Table 3 and the 1925–75 mean of the same source; Khazara 2014–18 from the MoW decadal record (config/ss_gate_bands.json).

4 What is wrong, and what 1968 says

Where the Soviet names are

Where the Soviet names are

Finding. Numbered triangles are the Soviet gauges, nodes and reach boundaries, from the Ravathoji headworks (1) to Khazara (12); lettered canals are those we could match to a Soviet asset with its 1951–70 flow; T is the Talligulyan spill, the return the modern gauge chain does not hold. The key at the right carries every name and number, so the map itself stays readable.

What it means. The same frame, markers and key are used for every map in this review, one map per place the text names.

Source: Geometry: zarafshan-water-body-delineation register (outputs/review-map/candidate-basin-clipped.geojson); nodes from the MoW station file, Tugainy and Jar-Osty by river kilometre (Гидрологическая изученность 1967, Table 3; straight-line position assumed); flows from Ресурсы поверхностных вод 1976, Table 4 (read); Talligulyan from the USAID/ZBUIS linear scheme (Q = 250 design), drawn as a straight line; settlements from OSM.

4 What is wrong, and what 1968 says

What the Soviet numbers change in the model

Every row is read from the 1971 volume unless noted; the model column is the durable export or the September test version. The full cross-walk with page references is in the working record.

TermSoviet sourcesModel todayWhat follows
Aquifer thicknessfan 1 000–1 200 m (70 m clean gravel over denser gravel); 40–50 m below the Karadarya–Akdarya confluence; piedmont 30–300 muniform 150 mvariable bottom; a partly sealed layer under the fan
Permeability Kfan 100–200 m/d; terrace III 12–17; piedmont 5.7; cover loam 0.2–0.431.6 / 5.5 / 2.1 m/d from literature meansraise the fan K three to six times
Seasonal swing of the water tablefan 13–18 m; valley 1–2.5 m; terrace III 0.8–1.3 m; Nuratau piedmont 0.1–0.2 m0.04 m simulatedstorage targets by unit for the month-by-month run
Precipitation recharge0.4 m³/s (0.5 % of inflow); 4–5 % in the oases10.9 m³/s (Rc = 0.10)Rc at most 0.02 on the valley floor
Irrigation and river losses79.4 m³/s; oasis norm 350–450 mm a year; canals : fields about 55 : 4530.5 m³/s; canals : fields 4 : 1larger fan losses, more field recharge
Inflow from the mountains1.7 m³/s underground; 8.7 m³/s emerge as springs and sais from the Zarafshan range; northern piedmont carries 3.421.7 m³/s injected; southern edge closedsmall underflow, sais and canals as surface recharge
Outflow to the west0.5–0.6 m³/s; 1.5–1.9 across the Khazara section (1949–1980)1.8–2.3 m³/s; band 0.5–2.0thin the boundary; aim at the lower half of the band
Pumping4.3 m³/s in 1968; 82 m³/s exploitable only with 65–97 m drawdown; sealed gravel at 50–200 m, older rock at 70–500 m37.8 m³/s from the register, 30–140 m conelower shallow abstraction or a deep layer for the well fields

Sources: Гидрогеология СССР т. XXXIX pp. 174–180, 260–273, 326–329, 341; Ресурсы поверхностных вод СССР т. 14 вып. 3 (1971) pp. 44–45, 91–92; Rubinova et al. 1988; Juraev et al. 2021. Cross-walk: docs/soviet_literature_crosswalk_2026-09-15.md §3.

4 What is wrong, and what 1968 says

Where the error sits

Where the error sits

Finding. 153 of the 381 observation wells, forty percent, are single-measurement wells in the northern foothill belt outside the mapped river deposits. Their typical error is 20 m too low and their spread the widest of any group. Wells in the river deposits and the terraces are centred near zero.

What it means. The model was built for the valley aquifer. It is being judged, to a large degree, on wells in a different unit that it represents with borrowed properties and no recharge.

Source: candidate well errors joined to calibration_targets_steady_state.csv (map_geology); diagnosis in the project issue log (GAPS.md).

4 What is wrong, and what 1968 says

Samarkand: the cone the model predicts, and the wells we know of

Samarkand: the cone the model predicts, and the wells we know of

Finding. If the register's pumping came from the shallow aquifer, the water table south-west of the city would fall by up to 140 m. The monitoring wells, screened at about 30 m, show no such drop. Half of the production wells the registry lists near Samarkand are screened in older older rock rock at 90 to 165 m.

What it means. Either the well fields draw on a deeper layer the model does not have, or far less is pumped from the gravels than the register says. The well fields themselves are not in any data we hold; their depths and metered volumes are the first things to ask for.

Source: heads of the May candidate minus its no-pumping control (scratch/ss_rapid_head_scatter_index); wells from data/momigeo/Samarqand_kuzatuv_q_malumotlari.xlsx (registry sample, not the utility well fields).

4 What is wrong, and what 1968 says

2 Northern foothills: the wells the model cannot place

2  Northern foothills: the wells the model cannot place

Finding. 153 wells, forty percent of all we have, sit outside the mapped river deposits on the northern piedmont, most of them tagged older rock in the registry and many stacked several to one model square. The model treats those squares as gravel with almost no recharge and puts the water table 20 to 100 m too low.

What it means. Something keeps the water high there that the model lacks: inflow from the hills, seepage from the northern canals, or a separate shallow system. The experts can say which. Until then these wells cannot be calibration targets for a valley-gravel model.

Source: aquifer_properties.npz zone_grid; calibration_targets_steady_state.csv (map_geology, registry age); regional_ss_workspace heads; northern tributary wells NB05a.

4 What is wrong, and what 1968 says

3 Where the river's water is taken out, and where it comes back

3  Where the river's water is taken out, and where it comes back

Finding. The delineation register shows what the model does not have: the river splits into the Ak-Darya and Kara-Darya around the Miankal island and the model follows only the Kara-Darya; below the headworks the gauged canals and the reservoir intake take another 108 m³/s that the model never removes. The balance sheet from the gauges says the river is down to 5 m³/s below the Narpay works, receives 14 from the returning Ak-Darya and 22 from sources no gauge records before Ziatdin, feeds Karmana, gains 14 more, and leaves 23 at Khazara.

What it means. About 36 m³/s enter the lower river unseen by any gauge we hold: collector returns, groundwater, or unmetered branches. Together with the Ak-Darya that is the water supplying Karmana and Khazara, and it is what the model must put back in the right place. The two new test version intakes near Navoi sit just above the Navbakhor gauge and are not in our records at all.

Source: zarafshan-water-body-delineation external data root, outputs/review-map/candidate-basin-clipped.geojson (2026-09-11); MoW decadal record 2014–2017 (mow_data_samarqand_long.csv) and station positions (mow_stations_swb.gpkg); model river cells regional_river_cells.json.

4 What is wrong, and what 1968 says

Irrigation, canals and drains: the recharge nobody has measured

Irrigation, canals and drains: the recharge nobody has measured

Finding. Two thirds of the valley floor is irrigated. The model assumes that 90 percent of canal losses and half of field losses seep to the aquifer, and that a network of drains two metres deep carries the surplus back to the river. Those three numbers decide both the fit to the wells and the river outflow, and none of them has been measured here.

What it means. One gauged drain, one measured canal loss, or one map of waterlogged land would constrain more than any further model run.

Source: regional_irrigation_fraction.npz (RS crop maps); OSM waterways; May candidate DRN-to-MVR cells; ledger docs/steady_state_water_balance_ledger.md.

4 What is wrong, and what 1968 says

What is missing in the data

Pumping

  • Coordinates and rates of production wells for Samarkand and Navoi; the register has them only for Bukhara
  • Annual abstraction series 2010–2025 by district; only the 2026 status exists
  • Piped exports to Navoi and Bukhara and sewer returns
  • Agricultural and unregistered pumping: no data at all

Aquifer geometry

  • Bedrock depth: no map; the model uses one uniform thickness
  • Filter depths give only lower bounds
  • No pumping-test data for the foothill belt and the oldest deposits

Surface water

  • Reliable river flow below the Ravathoji intakes only for 2014–2017
  • Canal deliveries per system, not per field
  • Khazara river flow used as a yearly average; monthly series needed for checking

Observations

  • 250 of 398 wells have a single measurement
  • Several wells share one 1 km square with levels differing by up to 74 m
  • 47 wells report levels above land surface: elevations need checking
  • No soil-moisture or tracer data for the soil between the surface and the water table

5 The plan

What the 1968 balance changes, and what we do with it

Until this month the model had nothing to be measured against except the wells. Every amount of water entering or leaving the ground was an assumption, and tuning could only trade one assumption for another.

The 1968 balance, the Soviet station records, Mamarasulov's 1972 book and the recovered Hydromet series change that. They say how much water leaks in and where, how much comes back and where, how thick the gravels are, and how far the water table swings each year.

Each of those numbers points at one thing the model does differently. So the plan is seven steps, each changing one thing, each with a test from the same sources, in the order the steps depend on each other. The first four need no new data.

5 The plan

Seven steps, one change each, one test each

Seven steps, one change each, one test each

Finding. Routing comes first because every later test reads the river; geometry before the well fields because a deep layer only makes sense on the Soviet thickness; recharge and reach targets before the edges because they fix the interior balance the edges have to close; storage last because it needs the month-by-month model on a settled long-term state. After the seventh, the coupled soil and canal system proposed in chapter 4 replaces the prescribed recharge, and the five Tashkent scenarios are run on it.

What it means. Each step is run through the same fixed set of checks, reviewed independently, and written down whether it passes or fails. Nothing enters the main model until it passes.

Source: Plan as of 2026-09-16 (assumed sequence; not yet run). Tests are the ratified gate bands (config/ss_gate_bands.json) and the Soviet ranges of docs/soviet_literature_crosswalk_2026-09-15.md §5. Workflow: Lean Environmental Modeling Orchestration 0.11 pinned in AGENTS.md.

5 The plan

Step 1 · Route the river: take the intakes out, put the returns back

Step 1 · Route the river: take the intakes out, put the returns back

Finding. The September test runs already tried the first half: with the gauged intakes removed at their nodes, the river runs dry below Narpay (red line) because the model has nothing that puts water back. The Soviet station volume names what did: the Talligulyan spill alone returned 32.7 m³/s into the Kara-Darya, the Bulungur and Miankal-Khatyrchi returns another 13, and today's collector surveys give 17–19 m³/s from the Samarkand mains and 9–11 from the Navoi mains. Step 1 adds these returns as gauged inflows at their reaches, before any recharge is touched.

What it means. Test: Khazara outflow inside the agreed 15–28 m³/s range with the river profile passing through the six observed points; delivered intakes at least 80 % of requested. If the returns are needed to hold the profile, the drainage cascade is physically required and step 4 targets it.

Source: Intakes: MoW decadal record 2014–17 means, GAPS.md §5 Iteration 6 (read). Returns: Ресурсы поверхностных вод 1976 Table 4 (1951–70), Mamarasulov 1972 p. 28, Shodiev & Chembarisov 2021 (docs/soviet_literature_crosswalk_2026-09-15.md §6.2, §10, §11). Profile: Iteration 6 result table (GAPS.md §5). Return positions are placed at the register geometry of the named canal or at the reach midpoint (Navoi collectors, Samarkand collectors: assumed).

5 The plan

Step 2 · A bottom that follows the Soviet thickness, two layers under the fan

Step 2 · A bottom that follows the Soviet thickness, two layers under the fan

Finding. The model is one layer, 150 m thick everywhere. The 1971 volume describes 1 000 m of gravel under the Samarkand fan with a 70 m clean unit on top, 8–60 m under the terraces, only 40–50 m below the confluence of the two branches, and 30–300 m of proluvium under the piedmonts. Step 2 builds a bottom surface from these statements, splits the fan into a clean upper layer at 100–200 m/d and a compacted lower one, and thins the lower valley to what it is.

What it means. Test: the error at the checked wells does not worsen and the fan cone of the September tests disappears at the same pumping; the western outflow moves toward the Soviet 0.5 m³/s without touching the boundary. The May ladder showed that a thicker uniform layer costs 3 m of error; a thinner variable one is the untested direction.

Source: Thickness statements: Гидрогеология СССР т. XXXIX pp. 174–177, 431–433 and Table 8 (docs/soviet_literature_crosswalk_2026-09-15.md §2.1–2.2, read). Class boundaries on the map are assigned from the model's geology zones (data/model/aquifer_zone_properties.csv), a fan radius of 32 km and the Khatyrchi longitude (assumed); the map is a proposal for the bottom surface, not a measured one. Today's thickness: regional_model_setup_v2.json.

5 The plan

Step 3 · Recharge where the Soviets measured it

Step 3 · Recharge where the Soviets measured it

Finding. The model and the 1968 balance take in about the same water, but from different places. The model recharges 10.9 m³/s from rain where the Soviets found 0.4, injects 21.7 m³/s at the edges where they measured 1.7 of underflow, and counts 30.5 m³/s of canal and field losses where they booked 79.4 for all surface water. Step 3 caps rain recharge at 2 % on the valley floor, applies 350–450 mm a year on the land actually irrigated, lets the river lose water through a very leaky river bed on the fan, and brings the mountain water in as sai and canal recharge rather than injected underflow.

What it means. Test: every inflow term inside the 1968 range, the sum within 10 % of 81.5 m³/s, and the checks unchanged on closure and flooding. The freed rain recharge has to reappear as canal, field and river-bed losses, or the balance tells us the modern irrigation delivery is smaller than assumed.

Source: Model today: docs/steady_state_water_balance_ledger.md and GAPS.md:1523-1524 (September candidate, read). Soviet values: Гидрогеология СССР т. XXXIX Tables 8 and 13, p. 247; Mamarasulov 1972 Table 9 (docs/soviet_literature_crosswalk_2026-09-15.md §2.3, §2.7, §3.3, §9). Planned bars: 2 % of 337 mm on the active area; 400 mm on the 3 272 km² irrigated area (computed, 41.5 m³/s, shown as 42); river-bed 20 m³/s = growing-season channel loss (Mamarasulov Table 9); the rest read. Irrigated cells: regional_rcha_evt_arrays.npz (measured count).

5 The plan

Step 4 · The 1968 balance becomes the test: what comes back up, reach by reach

Step 4 · The 1968 balance becomes the test: what comes back up, reach by reach

Finding. Table 8 of the 1971 volume gives, for four reaches of the valley, how much groundwater came back to the surface: 16 m³/s at the fan toe, 20 between Chapan-Ata and Ishtykhan, 18 between Ishtykhan and Tashrabad, and 20 lost on the fan itself. The April model returns 1 to 5 m³/s per reach. Step 4 makes the four reaches zones with a net-exchange target each, places drain cells in the Jumabazar–Samarkand spring belt with the 32 m³/s Karasu discharge as their flux target, and adds the outflow split 62 : 19 : 0.5 between water coming back up, evaporation and westward underflow as a whole-model check.

What it means. Test: each reach within 30 % of its 1968 value, the split within 10 points, and Khazara still inside its band. This is the first time the drains have a target of their own instead of absorbing whatever recharge is prescribed.

Source: Targets: Гидрогеология СССР т. XXXIX Table 8 pp. 270–273 and Table 18 (read; docs/soviet_literature_crosswalk_2026-09-15.md §2.4, §2.6); spring belt pp. 176, 326. Reach positions: MoW stations, Tugainy by river km (assumed), Ishtykhan from OSM. Model bars: SFR exchange of the April export (data/model/regional_ss_workspace/zrb_regional_ss.cbb) summed over the reaches (measured here); the candidate's drain return is not included.

5 The plan

Step 5 · The well fields: a deep layer, a pumping bracket, explicit exports

Step 5 · The well fields: a deep layer, a pumping bracket, explicit exports

Finding. The Soviet description resolves the pumping puzzle in principle: sealed gravel at 50–200 m and a a water-bearing layer in the older rock at 70–500 m exist under the fan, more than half the registered production wells near Samarkand are screened in the older rock beneath the gravels, and the 1971 forecast already said that pumping 21 m³/s from the shallow fan needs 65 m of drawdown. Step 5 puts the well fields into the lower layer of step 2, brackets their pumping between the 1968 use, the population-based estimate and the 2026 register until MMIG's annual series arrives, and routes the Damkhodja export and the Siyob sewage explicitly instead of returning them locally.

What it means. Test: no cone deeper than 5 m at the fan monitoring wells for any value in the bracket, the spring-belt flux still within its step-4 target, and the errors within 20 km of Samarkand losing their one-sided lean. Whichever pumping value passes is the one the shallow observations can support; the rest must come from depth or is not pumped.

Source: Wells: data/momigeo/Samarqand_kuzatuv_q_malumotlari.xlsx (registry sample within 34 km of the fan centre, counted here). Deep aquifers: Гидрогеология СССР т. XXXIX pp. 176–178; Ресурсы поверхностных вод 1971 p. 44; Juraev et al. 2021 (crosswalk §2.1, §6.4). Damkhodja and Siyob: Juraev et al. 2021 (crosswalk §6.4). Pumping values: Table 8 (4.3, 21.2), NB07a export (3.4), register (37.8, ledger row 99); cone 30–140 m: GAPS.md §5 Iteration 4. Arrow paths are schematic.

5 The plan

Step 6 · The edges: the northern piedmont as its own flow system, a thin western outlet

Step 6 · The edges: the northern piedmont as its own flow system, a thin western outlet

Finding. Forty percent of the observation wells sit on the northern piedmont, outside the geological map of the young deposits, and the model places them 20–100 m too low. The 1971 volume describes that belt as a separate body of groundwater in the foothill deposits: 300 m thick under loess, permeability about 5.7 m/d, 3.4 m³/s of natural resources entering along the Nuratau foot, flowing toward the valley at a steep angle. Step 6 gives the belt its own zone and a distributed inflow, brings the southern mountain water in as sai recharge with only 1–2 m³/s of underflow, cuts the eastern injection to a few m³/s, and thins the western outlet to the 40–50 m of alluvium that is there.

What it means. Test: the northern wells' systematic error closes to within 10 m or the wells are declared a separate aquifer and leave the set of wells the model is tuned to; western outflow inside 0.4–1.0 m³/s. Either outcome is a result: it decides whether those 153 wells constrain this model at all.

Source: Northern system and moduli: Гидрогеология СССР т. XXXIX pp. 177–178, 326, Table 8 unit 1 (crosswalk §3.5, §4 Q9); western outflow Tables 13, 18 and Rubinova et al. 1988 (crosswalk §3.4, §6.3); southern range Table 8 (crosswalk §2.7). Model today: GAPS.md:1598-1602, config/ss_gate_bands.json, docs/western_boundary_darcy_outflow_analysis.md (read). Wells: data/observations/calibration_targets_steady_state.csv, map_geology = Outside (counted here). CHD cells: regional_model_setup_v2.json.

5 The plan

Step 7 · Month by month: seasonal amplitudes by unit as the storage targets

Step 7 · Month by month: seasonal amplitudes by unit as the storage targets

Finding. The month-by-month model swings a few centimetres where the wells swing a metre, and the May sensitivity runs could not fix that with storage alone. The Soviet regime observations give the missing target by unit: 13–18 m a year on the fan, 1–2.5 m on the valley terraces, 0.8–1.3 m on terrace III, 0.1–0.2 m on the Nuratau piedmont. Step 7 re-runs the month-by-month model on the long-term state that passed steps 1–6, tunes storage by unit to those ranges, and then tunes jointly to levels, seasonal swing and the monthly Navoi and Khazara flows.

What it means. Test: median simulated swing within a factor of two of the observed one in every unit, and the fan swing reproducing the Soviet order of magnitude under the Soviet canal deliveries of step 3. Only after this do the Tashkent scenarios have a model that answers in the right season.

Source: Soviet amplitudes: Гидрогеология СССР т. XXXIX pp. 176, 328–329, Fig. 61–62; Ресурсы поверхностных вод 1971 p. 45 (crosswalk §2.5, read); 1930s drawdown Mamarasulov 1972 p. 29 (§10). Observed and modelled swings: data/model/transient_residuals.csv, yearly max–min per well with at least 10 months, averaged over years, median by map_geology (measured here). Units on the map are the model's geology zones, not the Soviet units; the assignment is approximate.

5 The plan

Each step, its test, and what happens if it fails

Every test is a number the deterministic checks or the Soviet sources supply before the run, written into the record first. A failed test is recorded and landed; it is not tuned away.

StepChangesMust passIf it fails
1 Riverintakes out, returns inKhazara 15–28 m³/s; profile through six gauges; intakes ≥ 80 % deliveredthe return list is incomplete: ask Uzhydromet for the collector series before step 4
2 Geometryvariable bottom, two fan layers, fan K 100–200error at the checked wells no worse; fan cone gone; west outflow toward 0.5keep 150 m; carry the deep layer only under the fan
3 Rechargerain 2 %, 350–450 mm irrigated, fan river bed, saisinflow terms in 1968 ranges; sum within 10 % of 81.5modern deliveries are smaller than assumed: revisit the intake accounting of chapter 2
4 Reachesfour zones with targets; spring-belt drainseach reach within 30 %; split 62 : 19 : 0.5 within 10 pointsthe drains are in the wrong place: move them to the mapped collectors
5 Well fieldsdeep layer, pumping bracket, exportsno cone > 5 m; belt flux held; Samarkand systematic error goneshallow pumping is at the low end of the bracket; the rest is not from this aquifer
6 Edgesnorthern system, sai recharge, thin westnorthern systematic error < 10 m; west 0.4–1.0 m³/sthe northern wells leave the set of tuning wells as a separate body of groundwater
7 Storagestorage by unit, joint tuningswing within ×2 by unit; fan order of magnitudestorage is not the limit: the canal signal is missing and only the coupled system supplies it

Bands marked provisional in config/ss_gate_bands.json remain owner-ratifiable thresholds; the Soviet ranges are read from the 1971 volume, not accepted science.

5 The plan

What each step needs from MMIG, Uzhydromet and Uzbekhydrogeology

Steps 1 and 4 · river and drains

  • Uzhydromet: monthly Khazara, Ziatdin and Navbakhor series 2010–2024 (Navoi 1967–2020 already in hand)
  • Provincial reclamation expeditions: collector discharges where they join the river, 2010–2024, for the Samarkand and Navoi mains
  • MMIG or ZBUIS: whether the Talligulyan spill still runs and where it enters

Steps 2 and 6 · geometry and edges

  • Uzbekhydrogeology: the six map appendices of the 1971 volume (hydrogeological map, natural and exploitable resources, reclamation zoning)
  • Shevchenko 1958 and Mirzaev 1974 on the zoning and reserves of the Zarafshan plain
  • Any bedrock-depth or geophysical section across the lower valley

Step 5 · well fields

  • MMIG: annual abstraction 2010–2025 by well field and district; screen intervals of the main fields
  • Volumes piped from Damkhodja to Navoi and Bukhara
  • Levels in the deep production wells over the same years

Steps 3 and 7 · recharge and storage

  • Zarafshan hydrogeological station yearbooks (from 1959): regime and balance observations, collector hydrographs
  • Canal deliveries by system and season 2010–2024 from the basin administration
  • Station volumes for Dupuli and the small tributaries (Urgut, Sazagan, Amankutan, Tusun)

6 Conclusions

Conclusions

Almost all groundwater in the valley is water that leaked from the river, the canals and the fields. The 1968 balance measured it; the model does not yet reproduce it.

The model's groundwater moves well under half the water it should and returns less than a tenth of it to the surface. That single shortfall explains the dry river at Khazara, the missing water between Narpay and Ziatdin, the too-small seasonal swing and the falling storage.

The model itself is sound, fast, documented and handed over. Its remaining errors sit in five places on the map, and each has a number from 1968 to be measured against.

Seven steps, each with a map, one change and one test, bring the model to the 1968 shape. The first four start now on existing data; the last three need the pumping records, collector flows and map appendices that MMIG, Uzhydromet and Uzbekhydrogeology hold.

Once the seven steps are passed, the scenarios agreed with MMIG can be run on a model that carries the right water in the right places.

6 Conclusions

A working platform, ready for its next stage.

hydrosolutions GmbH · Zurich · September 2026

7 Appendix

Appendix

A The coupled soil, canal and groundwater system, with the Chu basin as the worked example

B What was agreed with MMIG in Tashkent: the scenarios and the earlier list of improvements

C The questions put to the experts on 16 September 2026

D Method details: population and rainfall, tributaries, the balance remainder, model geometry and workflow, sensitivity runs, hydrograph variants, storage tuning, how each step is run

E The Soviet sources: what was consulted, what was taken from the shelf, and whether 1968 is worth modelling

7 Appendix

A · The missing link: the soil between canal and aquifer

In this valley the aquifer is fed mainly by irrigation. How much of the diverted water seeps down, how much evaporates and how long it takes to arrive depends on soils, crops, canal lining and the state of the water table. The present model replaces all of that with fixed fractions.

The setting tests and the step-by-step concept tests both point at these fractions as what controls the fit to the wells and the river outflow. No amount of tuning the aquifer settings can compensate for a wrong recharge input.

Simulating the land surface and the canal system explicitly, and letting the aquifer model receive recharge from them, is the natural next stage. This is what was proposed and implemented for the Chu River basin plan.

7 Appendix

A · How it was done in the Chu basin: one water cycle, three engines

A · How it was done in the Chu basin: one water cycle, three engines

Finding. In the Chu basin plan the natural hydrology (SWAT+), the managed canal network (TaqSim) and the valley aquifer (MODFLOW 6) are three engines that hand water to each other every day. Water seeping through the soil and canal losses become groundwater recharge; river gains and drain returns flow back to the river.

What it means. The Zarafshan valley has the same anatomy: mountain inflow, a diverted river, canals, irrigated plain and a shallow aquifer. The pattern transfers directly.

Source: Chu River Basin Plan 2027–2031 modelling showcase, hydrosolutions 2026; figure reused unchanged.

7 Appendix

A · The Chu water system as one picture: which engine owns what

A · The Chu water system as one picture: which engine owns what

Finding. Snow and glacier melt, the reservoir, canals and irrigation, pumping, recharge, groundwater flow and the springs that return water to the river, drawn as one block of landscape. The brackets show the reach of each engine and the coupling zone where they hand water to each other.

What it means. Replace the Kyrgyz Ala-Too with the Zarafshan headwaters and the Chu with the Zarafshan, and the picture is the one this valley needs.

Source: Chu River Basin Plan 2027–2031 modelling showcase, hydrosolutions 2026; illustration reused unchanged.

7 Appendix

A · How the Chu engines are wired together

A · How the Chu engines are wired together

Finding. Every transfer between the engines is booked once, with source, receiver and date, and the groundwater account closes to a tiny remainder. The aquifer model receives simulated recharge instead of a prescribed array, and the river receives simulated gains and losses.

What it means. The linking code, the accounting checks and the tuning workflow exist and are documented. Reusing them for the Zarafshan is an engineering task, not new research.

Source: Chu River Basin Plan 2027–2031 modelling showcase, hydrosolutions 2026; figure reused unchanged.

7 Appendix

A · Proposed next stage for the Zarafshan

A · Proposed next stage for the Zarafshan

Finding. A land-surface and soil model for the irrigated plain and the mountain catchments, an allocation model of the Ravathoji diversions, canals and reservoirs, and the existing MODFLOW 6 model as the aquifer engine. Recharge, water loss to the air and river exchange become computed exchanges.

What it means. The regional model built so far is not lost: it becomes the groundwater engine of the coupled system. The long-term average work identified exactly the inputs the linked system would replace.

Source: Pattern: the SWAT+ – TaqSim – MODFLOW 6 coupling of the Chu basin plan (hydrosolutions, 2026); engines for the Zarafshan are a proposal, not an implementation.

7 Appendix

B · What was agreed with MMIG in Tashkent, May 2026

B · What was agreed with MMIG in Tashkent, May 2026

Finding. The working session fixed five management scenarios and their combinations, the output table each run has to deliver, and two data points on pumping: the population-based estimate is only the drinking and industrial part, and agricultural pumping needs MMIG's own data. It also flagged the systematic bias between simulated and observed levels as the first problem to solve.

What it means. The scenarios are the questions the model must answer; the bias diagnosis in this review is the work that makes the answers trustworthy.

Source: Whiteboard of the Tashkent working session with MMIG, May 2026 (photograph).

7 Appendix

B · Five scenarios and their combinations

Each run reports river inflow, river outflow, the piezometric water level from the model and its change against 2024. In the long-term average runs inflow equals outflow; in the month-by-month runs the difference is the change in the water stored in the aquifer.

ScenarioDriverChange tested
S1River and climateRiver inflow at Ravathoji reduced from 140 to 100 m³/s
S2PopulationPopulation of about 3 million grows by 20 %; domestic and industrial pumping scale with it
S3AgricultureGroundwater pumping for irrigation reduced by 40 m³/s
S4ClimatePrecipitation halved
S5ClimateEvaporation demand of the air increased by 25 %
S1 + S2CombinedLess river inflow and more people
S1 + S2 + S3CombinedLess river inflow, more people, less irrigation pumping

Not yet run: under the project workflow, scenario runs start once the model passes the checks of the seven steps in chapter 5.

7 Appendix

B · The improvement list of May 2026, superseded by chapter 5

1 Settle the pumping record

  • Obtain the annual abstraction series 2010–2025 and well-field coordinates from MMIG
  • Bound the pumping over the model period between the 2010 and 2026 registers
  • Route well-field exports and city sewage explicitly instead of local return

2 Clean the set of observation wells

  • One value per model square, weighted by well count and spread
  • Exclude edge and above-ground wells; separate static from time-series wells
  • Report all metrics on the cleaned and the full well set

3 Fix the foothill edges and the river

  • Recharge from the mountain front and a zone for the older rock along the northern edge, or a separate aquifer for those wells
  • Split river sections at the square boundaries
  • Check river outflow against monthly Khazara flow

4 Link the soil and canal system

  • Set up SWAT+ for the plain and the mountain catchments and an allocation model for the canals
  • Pass recharge and river exchange to the existing groundwater model daily, as in the Chu basin
  • Only then: joint tuning to levels, seasonal rise and fall, and river flow

Chapter 5 turns this list into seven iterations, each with its map, its change and the test it must pass.

7 Appendix

C · 1 The pumping puzzle: what we need to ask about the well fields

The register says the well fields around Samarkand pump about ten times more than the population would explain. Put into the model, that pumping drops the water table under the city by 30 to 140 m; the measured levels show no such drop. Extra inflow from the mountains, higher permeability and leaking pipes do not close the gap. Either the water comes from a deeper layer the model lacks, or far less is taken from the shallow gravels than the register says. The observation wells are only about 30 m deep and may not see a deep well field at all.

  • How deep are the screens of the main well fields near Samarkand (Chupan-Ata, Dargom terrace, Damkhodja): shallow gravels, a deeper layer, or both?
  • How much did these well fields actually pump each year from 2010 to 2024? Is the register a licence, a design capacity, or a metered amount?
  • Where does the water go: how much is piped to Navoi and Bukhara, how much serves Samarkand, how much returns through leaking mains or the Siab collector?
  • Have levels in the deep production wells fallen over the last fifteen years, and do the shallow monitoring wells nearby show the same trend?
  • Is there a known groundwater inflow from the mountains south of Samarkand, and does anyone have a figure for it?

Written before the meeting of 16 September 2026; places are shown on the maps of chapter 4.

7 Appendix

C · 2 Questions about the northern foothills

  • Are the wells north of the river, in the Neogene foothills, in the same aquifer as the valley gravels, or in a separate, perched system on the piedmont?
  • What keeps their water levels high: inflow from the hills, springs, old kariz galleries, or seepage from the canals of the northern irrigation zone?
  • Do these wells rise and fall with the seasons, or hold steady? Steady levels would point to a slow, deep source.
  • Where do the three northern streams (Oqsoy, Tusunsoy, Qorasuv) actually sink into the ground, and has anyone measured how much of their flow reaches the valley?
  • Should the piedmont be treated as its own zone with its own permeability, or should these wells be left out when the valley model is tuned?

Numbers refer to the overview map.

7 Appendix

C · 3 Questions about the river and Khazara

  • Below the Narpay works the river carries only about 5 m³/s (2014 to 2017), yet at Ziatdin it carries 41 again. The Ak-Darya brings 14 of that. What brings the other 22: the Taligulyan collector, other drains, springs, or a branch we do not have? And what adds another 14 between Karmana and Navbakhor?
  • In a normal year, how much water leaves the valley past Khazara: the gauge's 15 to 28 m³/s, or more? Is the 2014 to 2018 record complete, and is it typical of 2010 to 2024?
  • Which canals take water directly from the main river below Samarkand, and how much? The gauges say Narpay 25 m³/s and the two Karmana channels 29 m³/s. Are there others, for example the Navoi pumping station at 40 m³/s, and do they run all year?
  • Where is the river known to gain groundwater, and where does it lose water to the ground? Are there gauges on the river between Ravathoji and Khazara?
  • Are the collector drains gauged where they join the river? How much drainage water reaches the river in a year, and does any of it enter below the Khazara gauge?
  • Of the 76 m³/s taken for irrigation, how much comes back to the river on the surface, through drains and spillways, rather than through the ground?
  • Are there other withdrawals from the river between Samarkand and Khazara that we do not have in the records, such as pump stations or small canals?

Numbers refer to the overview map.

7 Appendix

C · Questions about canals, fields and drains

  • How much of the canal water is lost before it reaches the fields, and which of the main canals are lined?
  • How deep are the collector drains, how dense is the network, and is it still maintained and flowing?
  • Is drain flow measured anywhere in the valley? Even one gauged collector would tell us how much irrigation water comes back through the ground.
  • Where is the water table so shallow that fields are waterlogged or salinised, and is there a map of that?
  • How much groundwater is pumped for irrigation from private wells in the irrigated belt, and is any of it metered?

These questions have no number: they apply to the whole irrigated valley floor.

7 Appendix

C · 4 5 The edges, the register and the farm wells

  • 4 Southern mountain front. Is there a known underground inflow from the Zarafshan range south of Samarkand, on the Urgut side, and what feeds the springs there?
  • 5 Western edge. After Khazara, does groundwater keep flowing west under the valley floor towards Navoi, or does it all surface? Are there wells or gauges that show it?
  • The register. What does “отбор” mean in practice: a licence limit, a design capacity, or a measured volume? Can we get the yearly series 2010–2025 and the well-field table (табл. 2) for Samarkand and Navoi?
  • Farm wells. How many private irrigation wells are there in the valley, how deep are they, and is any of that pumping metered?

Numbers refer to the overview map.

7 Appendix

D · More people, more pumping

The model area is home to almost five million people. Their number grew by a third between 2010 and 2024, and domestic demand in the model grows in step.

Even the population-based estimate, which is the lower bound, rises from 0.09 to 0.12 km³ a year over the simulation period.

D · More people, more pumping

Source: data/population/yearly_pumping_summary.csv (WorldPop 2020 scaled by national growth).

7 Appendix

D · Dry years come in runs

Rain and snow over the valley are the only recharge that does not depend on the canals. Three of the last five years were well below the 2010–2024 mean.

Recharge from rainfall is small in absolute terms, about a tenth of what falls, but the same dry years also cut the river inflow that feeds the canals.

D · Dry years come in runs

Source: CHIRPS v3 monthly rasters averaged over the model outline (measured in this review); the 10 % share of rain that recharges is the water-balance assumption.

7 Appendix

D · Three ungauged tributaries from the north

Oqsoy, Tusunsoy and Qorasuv have no gauges. Their flow was estimated from 85 similar basins with gauges, scaled by elevation and checked against a published estimate for Tusunsoy.

They add 4 m³/s on average, mostly in March and April, and enter the model as groundwater inflow along the northern edge because their water seeps into the ground on the fans.

D · Three ungauged tributaries from the north

Source: data/mow/northern_tributary_inflows.csv.

7 Appendix

D · Where could the 1.4 km³ a year go?

Rain is not part of the remainder: its evaporation is defined as rain minus the assumed 10 % that recharges. The remainder is diverted river water that does not show up as crop water use; over the long term it must evaporate elsewhere, leave the valley unmeasured, or be error in the large gauged terms.

Possible explanationPlausible size, km³ per yearBasis
Evaporation from shallow groundwater and wet ground: canal banks, waterlogged fields, drains, riverside vegetation, salt flats. The usual fate of canal seepage.up to about 1model: 0.21 realised against 2.06 potential
Crop water use under-counted by the satellite product, which covers mapped cropland only: orchards, household plots, urban greenery are outside it.0.2 to 0.4cross-check ratio 1.04 to 1.16; assumed
Register pumping that is consumed or piped out of the valley: the three well-field districts report 0.71 a year, only 0.13 is in the balance.0.1 to 0.5register totals; export share unknown
Drainage water leaving the valley unmeasured, to Kashkadarya or below the Khazara gauge.unknown, possibly several tenthsno drainage outflow data
Rainfall over 10 200 km² from a satellite product; a 10 % bias is 0.34, and the 10 % recharge share is an assumption.± 0.3assumed
River gauging: 10 % on the Ravathoji inflow is 0.46; outflow and transfers rest on 2014–2017 canal records.± 0.3 to 0.5assumed
Change in groundwater storage and underflow past Khazara.about 0.1 togetherlevel trends; Darcy estimate

Honest reading: perhaps half is real recharge that leaves again as evaporation from a shallow water table and through drains; the rest is a mix of satellite under-coverage, unrecorded pumping exports, drainage outflows and gauge error. The 0.96 / 0.46 split on the previous slide is an assumption, not evidence.

7 Appendix

D · The same valley, balanced independently: Ragettli, Kreiner et al. 2025

Their Samarkand oasis (12 390 km², outflow at Navoi, 2017–2022) overlaps our model area (10 211 km², outflow at Khazara, 2014–2017); they book losses, winter leaching and pumping explicitly, so their residual is only 0.1 to 0.2 km³.

Term, km³ per yearThis reviewRagettli et al.Comment
River inflow after transfers out3.254.26theirs includes Tusun and 12 small rivers, before the Tuya Tortar transfer
Rain: full amount / runoff only3.44 (0.34 recharges)1.08they count the 30 % that runs off
River outflow0.72 at Khazara; Uzhydromet Navoi 2010–2020: 0.801.32 at NavoiNavoi lies upstream; more is diverted below it; the Hydromet decade mean confirms our band
Crop water use from satellite1.291.88theirs adds orchards, gardens, parks, canal-side vegetation (0.57)
Canal and field losses1.38 gross, 0.96 to recharge2.01 as a usenot split into recharge and evaporation
Winter leaching, pre-season irrigationnot included0.78a recharge process our model lacks
Groundwater pumping0.13 drinking, industry0.96 ± 0.47 as inputsupports the register magnitude (0.87)
Unattributed remainder1.420.12 to 0.17absorbed by the explicit terms above

Losses 2.0 plus winter leaching 0.8, minus 1.0 pumped back for irrigation and 0.6 of consumptive use our satellite term misses, give about 1.2, close to our 1.4. Lessons: pumping of about 1 km³ a year is real; winter leaching recharge is missing; their Samarkand drainage volumes point to groundwater draining through the collectors, as our drain concept assumes.

7 Appendix

D · The model in three dimensions

D · The model in three dimensions

Finding. Seen as a block with the vertical stretched about 30 times, the model is a thin slab that follows the land surface down the valley: 150 m thick everywhere, 1 300 m above sea level at Ravathoji and 310 m at Khazara, over a length of 230 km.

What it means. The geometry is simple by design: one layer, one thickness. The bedrock depth that would give the slab its true shape is the largest structural unknown.

Source: data/model/regional_model_setup_v2.npz (top, bottom, active area); drawn in model-grid coordinates.

7 Appendix

D · From raw data to a working model

D · From raw data to a working model

Finding. Twenty processing steps in fixed order turn the raw data into model inputs, run the long-term average model with its tests and adjustments, then the month-by-month model, and hand everything over through an open code repository, a viewer and a workshop.

What it means. Every figure in this review was regenerated from those steps on 14 September 2026. MMIG staff can do the same.

Source: the project's documented processing chain.

7 Appendix

D · Which assumptions matter for the water levels

D · Which assumptions matter for the water levels

Finding. 5 000 combinations of model settings were run, 3 091 of them successfully. Five settings control the fit: how easily water moves through the upper and middle terraces, the evaporation rate from shallow groundwater, recharge from rain, and aquifer thickness. River-bed leakiness, the fixed edge levels, bedrock and channel roughness barely matter.

What it means. Water-level measurements cannot pin down the river settings at all; that needs river-flow data. And no combination brings the typical error below about 15 m.

Source: scratch/lhs_results.csv (test-run sample); influence ranking against the typical error in the young river deposits (measured here).

7 Appendix

D · No version beats the April model yet

D · No version beats the April model yet

Finding. The best of 3 091 tested settings brings the typical error down to 16 m, but with levels now too high on average. The set found by automatic tuning software, re-run in the full model, is worse than the April model at 27 m. The September candidate with the full register pumping sits at 25 m.

What it means. Tuning the settings has reached its limit. The remaining error comes from the model's structure and from the observation wells themselves, not from the setting values.

Source: the executed processing steps of 2026-09-14; scratch/lhs_results.csv; scratch/ss_rapid_v1_kdown_sfr_connectivity/scenario_metrics.csv.

7 Appendix

D · Step by step towards a more realistic concept

D · Step by step towards a more realistic concept

Finding. Five changes to the model concept were stacked on the April model in May 2026: canal intakes taken out of the river, canal and field losses counted in full with shallow drainage channels, a thicker aquifer, and the full register pumping with local return flows. Each step after the first worsened the fit to the wells, and the river outflow at Khazara stayed at 50 to 59 m³/s against an observed 23.

What it means. Two lessons: the excess river flow comes from the fully counted losses that the drainage channels return to the river, and the register pumping cannot be sustained where the model places it.

Source: scratch/ss_rapid_scenarios and scratch/ss_rapid_v1_kdown_sfr_connectivity scenario_metrics.csv (May 2026); observed outflow from the water-balance constraints.

7 Appendix

D · The same wells with the static offset removed

D · The same wells with the static offset removed

Finding. Shifting each simulated series so that it starts at the first observation separates two kinds of error. The constant offset, several metres to tens of metres, is a steady-state problem: geometry, permeability and recharge. What remains after the shift is the dynamic error: the direction of the multi-year drift is mostly right, the year-to-year response is visible, and the seasonal swing is far too small.

What it means. The two errors have different owners. The offset goes to the steady-state iteration now under way; the dynamics go to the storage coefficients of the transient model.

Source: data/model/transient_residuals.csv (run of 2026-09-14), months from 2012 onward (first two years of settling-in dropped). Shift = first observation minus first simulated value of each well.

7 Appendix

D · Standardised: the seasonal rhythm is there, only too small

D · Standardised: the seasonal rhythm is there, only too small

Finding. Centring each series on its own mean and dividing by its own spread removes both the offset and the amplitude and leaves only the shape. Where the two curves then move together, the model has the right timing of recharge and drainage and only the wrong storage coefficient; where they diverge, the process itself is missing. The legend gives the spread that was removed: observed swings of one to several metres against simulated swings of centimetres to decimetres.

What it means. This is the test the Soviet amplitudes make quantitative: 13 to 18 m on the fan, 1 to 2.5 m in the valley, 0.1 to 0.2 m on the Nuratau piedmont. A model whose standardised curves match can be brought to those amplitudes by the storage coefficients alone.

Source: data/model/transient_residuals.csv (run of 2026-09-14), months from 2012 onward (first two years of settling-in dropped). Each series centred on its own mean and divided by its own standard deviation over the plotted months; the SD in the legend is the amplitude the standardisation removed.

7 Appendix

D · What controls the seasonal rise and fall

D · What controls the seasonal rise and fall

Finding. Five hundred month-by-month runs varied the storage and flow settings together; 480 succeeded. How much water the recent river deposits and both terrace generations can release from storage controls the seasonal rise and fall, with a smaller opposite effect from how easily water moves through the middle terraces.

What it means. The published storage values are too high for this valley. Month-by-month data pin down what the long-term average model cannot see, so the next tuning must use both.

Source: scratch/lhs_transient_results.csv; influence ranking against the seasonal rise-and-fall error in the young river deposits (measured here).

7 Appendix

D · Register against population, district by district

D · Register against population, district by district

Finding. For the thirteen Samarkand districts with a purpose breakdown, the MMIG register reports 27.5 m³/s of drinking and industrial abstraction; population and per-capita norms explain 3.6 m³/s. The gap sits almost entirely in the three districts with regional well fields.

What it means. Well fields export water by pipe to Navoi and Bukhara and return city water through the sewer collector. Without well coordinates, annual series and export volumes, the model cannot place this pumping, and unregistered farm pumping is not in either number.

Source: data/population/modelled_vs_observed_comparison.csv (MMIG ОТБОР register).

7 Appendix

D · How each iteration is run, and what MMIG will see

One branch per iteration, opened with a written contract: the objective, the inputs by identity, what must not change, the test per stage and the conditions under which only the owner decides. The runs use the rapid steady-state scenario tooling of May, so a full ladder takes minutes, not days.

The deterministic gate runs on every scenario: convergence, budget closure, the routed-river identity, the Khazara and western bands, the cleaned-well error and the flooded-cell share. Only a configuration that passes gets an independent critique and is then promoted into the durable notebooks, which re-triggers the month-by-month run.

Each landed iteration adds one entry to the project record with the setup, the numbers and the diagnosis, whether it passed or not. MMIG receives the updated deck and viewer after each promotion, and the map on each step slide is redrawn from the new state.

Steps 1 to 4 can be run back to back on existing data. Steps 5 to 7 start as soon as the corresponding data arrive; step 7 also needs steps 1 to 6 landed. The coupled soil and canal system follows step 7 and is a separate proposal.

7 Appendix

E · The Russian-language sources newly consulted

E · The Russian-language sources newly consulted

Finding. Three Soviet syntheses (the 1971 Uzbek and 1972 Tajik volumes of Hydrogeology of the USSR and the 1963 Academy volume), four volumes of the Surface Water Resources series for the Amu Darya basin (the 1971 monograph, the 1967 station catalogue, the 1976 and 1980 station records), Mamarasulov's 1972 book on the water problem of the Zarafshan basin, Kudelin's 1966 underground-runoff summary, the 1975 methodology of regional groundwater assessment, the 1981 machine-irrigation scheme, and five papers from 1988 to 2025 on the basin's water balance, collectors and groundwater use.

What it means. Still to obtain: Mirzaev 1974 on the groundwater reserves of Uzbekistan, Shevchenko 1958 on the hydrogeological zoning of the Zarafshan plain, Reshetkina 1957, Khikmatov et al. 2016, and the six map appendices of the 1971 volume, all held in Tashkent.

Source: Bibliography: docs/references/soviet_zarafshan_sources.bib and .ris (31 entries incl. items cited but not yet obtained). Extraction: docs/soviet_literature_crosswalk_2026-09-15.md.

7 Appendix

E · What we take from the shelf into the next iteration

Ledger priors. The 1968 balance and the Kermine-Kenimekh balance enter the steady-state ledger as external rows; the split of the outflow between emergence, evaporation and westward underflow, 62 : 19 : 0.5, becomes a check the model has to pass.

Reach targets. The Soviet units, Ravathoji to Tugainy, Tugainy to Chapanata, Chapanata to Ishtykhan, Ishtykhan to Tashrabad, become zones with a target for net recharge minus discharge; the Karasu springs at the fan toe become a flux target of 16 to 32 m³/s.

Properties and structure. Fan permeability 100 to 200 m/d, a bottom that follows the Soviet thickness, a semi-confined fan layer for the well fields, seasonal amplitudes by unit as storage targets.

Recharge. Rain recharge capped at 2 percent on irrigated land; irrigation recharge of 350 to 450 mm a year; river-bed losses on the fan of the order of 60 m³/s at the flood peak through a high-leakance bed; mountain water as sai and canal recharge, not injected underflow.

Still to fetch. Mirzaev 1974 on the groundwater reserves of Uzbekistan, Shevchenko 1958 on the hydrogeological zoning of the Zarafshan plain, Mamarasulov 1972, the station volumes for Dupuli and the small tributaries, and the six map appendices of the 1971 volume, which are held by Uzbekhydrogeology.

7 Appendix

E · The 1968 balance is history. Is the pre-development state worth modelling?

What has changed since 1968. Pumping has grown from 4.3 m³/s on the fan to tens of m³/s across the valley; the collector network from about 1 800 km to more than 3 500 km; the Kattakurgan reservoir was enlarged in 1966 and the machine canals of the 1960s and 1970s put 391 thousand ha on guaranteed Zarafshan supply; the river below Ravathoji now carries about 36 m³/s against 165 to 190 at the mountain exit in the Soviet records. The 1968 numbers are a reference for structure and process, not a target for 2010 to 2024.

Why a 1968 steady state would still pay. Pumping was small then, so the 1968 balance isolates the natural system: river losses on the fan, permeability, bed leakance, the spring belt at the fan toe, drainage and evaporation. A model that reproduces Table 13 and the reach-by-reach emergence of Table 8 with 4.3 m³/s of pumping has its recharge and discharge machinery fixed; the 2010 to 2024 run then adds pumping, the enlarged canal and drain network and the measured river, and the difference between the two states is the human footprint the ministry asks about.

What it would take. Available: the 1968 balance and unit resources, water-table depths and amplitudes, the 1969 canal and collector lengths, mean river discharges, reservoir volumes. Missing: the 1968 hydroisohypse map (appendix 1 of the 1971 volume, held by Uzbekhydrogeology), irrigated area and canal deliveries by system for the 1960s, and the observation wells of the Zarafshan hydrogeological station from 1959. Until those are found, a 1968 run is a plausibility check on the model's processes, not a calibration.

1 / 1