Pakistan Projects 22% Rabi Water Shortage as Dam Delays Put Wheat Supplies Under Pressure

Winter crops face tighter supplies as reservoirs fall short

Pakistan's Indus River System Authority (IRSA) has approved a projected 22% irrigation water shortage for Punjab and Sindh during the 2026 to 2027 Rabi season, with reservoir storage entering October at 9.343 million acre feet, compared with 13.214 million a year earlier. The two provinces produce most of the country's wheat, making the allocation a concern for food supplies already strained by a disappointing harvest and rising flour prices.

The forecast covers October 1, 2026, to March 31, 2027. Provincial withdrawals are planned at 29.183 million acre feet (MAF), against roughly 38 MAF envisaged under the 1991 Water Apportionment Accord. Balochistan and Khyber Pakhtunkhwa are exempt from the supply cuts applied to Punjab and Sindh.

IRSA's assessment points to lower river inflows, insufficient stored water and construction restrictions at Tarbela Dam. Delays affecting its Tunnel 4 outlet and Tunnel 5 hydropower project have limited the ability to retain water for later use. These are distinct from Pakistan's political dispute with India over the Indus Waters Treaty.

The immediate concern is how reliably canals can supply farmers through planting and crop growth. More than 80% of Pakistan's agriculture depends on irrigation supplied by the Indus Basin. Wheat is the principal Rabi crop, alongside gram, lentils, barley, mustard, rapeseed and tobacco.

The forecast does not establish how much the next harvest will fall, or whether wheat imports will need to increase. It does establish that farmers are beginning the season with less stored water than expected, while consumers are already paying substantially more for flour.

What the 22% forecast measures

The IRSA Advisory Committee, chaired by IRSA Chairman Amjad Saeed, approved expected inflows of 21.268 MAF at rim stations. These are river measurement points used to assess water entering the managed irrigation system. An acre foot is the volume needed to cover one acre of land to a depth of one foot; MAF expresses that quantity in millions.

The committee assessed different scenarios using the WAA-Tool, developed with CSIRO Australia, the University of British Columbia Watershed Model and participating stakeholders. Its operating assumptions included conveyance losses of 4% in the Indus zone and zero in the Jhelum and Chenab zones. Conveyance losses are water lost while moving through the delivery system.

Expected river inflows and planned provincial withdrawals are different measures. Withdrawals can draw on reservoir stocks as well as incoming river water, subject to losses and operating restrictions. Subtracting the inflow forecast from provincial allocations would therefore not provide a separate, valid estimate of the shortage.

IRSA explained the approved forecast in its statement following the meeting:

Based on the agreed inflows and other system operational parameters, overall system shortage worked out at 22pc for both Punjab and Sindh, which was approved

The estimate can be reviewed during the season. Reports also describe internal assessments by Punjab, the Water and Power Development Authority (WAPDA), and IRSA that put the shortage at 26%. The lower approved figure reportedly followed objections from Sindh. That difference matters: 22% is the adopted planning estimate, not proof that all assessments reached the same result.

How the provincial allocations compare

Punjab's allocation is 15.564 MAF and Sindh's is 11.748 MAF. Together, those figures represent about 94% of the reported 29.183 MAF total. Balochistan is assigned 1.171 MAF, while reports differ slightly on the allocation for Khyber Pakhtunkhwa.

One figure for Khyber Pakhtunkhwa is 0.710 MAF; another is 0.701 MAF for its Chashma Right Bank Canal system. Using 0.701 MAF makes the four listed allocations total exactly 29.183 MAF. Using 0.710 MAF produces 29.193 MAF. The difference should not be silently resolved, although it is small beside the wider shortage.

The distribution uses a formula that applies shortages to Punjab and Sindh while exempting the other two provinces. Sindh has objected to that approach and sought allocation under Paragraph 2 of the 1991 Accord. The dispute concerns how insufficient water is shared, rather than whether the reservoirs contain enough to meet full requirements.

Planned withdrawals are about 1.6% above the ten year average of 28.73 MAF. That does not contradict the shortage forecast: the average records past withdrawals, while the shortage is assessed against seasonal requirements and allocation rules. Historically constrained deliveries can leave an average below the full entitlement.

A weaker summer left less water for winter

During Kharif, the April to September 2026 season, actual river inflows were 95.44 MAF. That was 8% below expectations, 7% below the ten year average and 22% below the previous year. River flows depend on rainfall, snow and glacier melt, and temperatures, so less incoming water can reduce both immediate irrigation supplies and the amount available for storage.

By September 30, reservoirs held 9.343 MAF against an expected 12.014 MAF, a deficit of 2.671 MAF, or about 22.2%. Storage was also about 6% below the ten year average of 9.938 MAF.

The decline from the previous year's 13.214 MAF was larger: 3.871 MAF, or about 29.3%. Some accounts describe the previous season's roughly 13 MAF as about 27% higher than current storage. That percentage does not match the detailed figures. Using 13.214 MAF, last year's stock was about 41.4% higher than this year's, while this year's stock was 29.3% lower.

Figures attributed to WAPDA's daily reporting show Tarbela holding 4.236 MAF, 18% below its five year average. Mangla held 5.095 MAF, within 3% of its average. The uneven position points to particular pressure at Tarbela, rather than an identical decline across the two main reservoirs.

Water passing downstream of Kotri during Kharif was 10.40 MAF, compared with 13.05 MAF expected, a ten year average of 15.43 MAF and 30.85 MAF the previous year. The latest figure was roughly one third of the previous year's volume. These flows form part of the water balance; they should not automatically be treated as water that could all have been stored or diverted.

Tarbela construction limits the storage response

More than 3 MAF reportedly could not be stored at Tarbela because WAPDA imposed filling restrictions during construction. Reports also describe increased releases from Mangla to compensate for Tarbela's constraints, making it harder to replenish Mangla before Rabi.

The projects in question are the Tunnel 4 Low Level Outlet and Tunnel 5 Hydropower Project. Their relevance extends beyond electricity generation: construction and reservoir operating restrictions affect how much water can be held back for later irrigation.

IRSA's committee urged early completion, saying the works were already substantially delayed and were contributing to provincial shortages during Kharif. Reports disagree on their original deadlines. One account places both deadlines in March 2025; another says Tunnel 4 and its outlet were due in 2022, while Tunnel 5 was due in July 2025.

No firm replacement completion date is established. One report describes delays potentially extending into early 2028 and says WAPDA indicated that full Tarbela capacity would be unavailable during the 2027 and 2028 Kharif seasons. That is a warning of continuing restrictions, not a confirmed commissioning schedule.

Summer releases add another allocation dispute

IRSA's Kharif review recorded provincial use below Accord shares by 11% in Punjab, 7% in Sindh, 28% in Khyber Pakhtunkhwa and 58% in Balochistan. Those seasonal totals sit alongside a separate record of additional supplies during late Kharif.

During that later period, Sindh used 1.262 MAF and Punjab 0.077 MAF as flood supplies under Paragraph 4 of the 1991 Accord, beyond the Paragraph 2 allocation referenced in the review. Sindh's additional volume was about 16 times Punjab's. Additional use during one part of the season can coexist with use below an entitlement across the season as a whole.

Reports allege that political intervention contributed to releases from storage and that promised adjustments to provincial shares were not completed before the season ended. The recorded volumes support scrutiny of allocation decisions, but they do not by themselves establish a political motive or quantify an avoidable economic loss.

The practical issue is the reserve left for winter. An accounting adjustment to a province's share cannot physically restore water already released. Assessing the decisions would require the expected inflows, storage targets, operating restrictions and available alternatives at the time.

Wheat risks arrive after a weak harvest

Pakistan's previous wheat harvest is estimated at 27.5 million tonnes, against a production target of 29.7 million tonnes and estimated domestic consumption of 31.9 million tonnes. Those figures imply a 2.2 million tonne shortfall against the target and a 4.4 million tonne gap against consumption. The two gaps measure different things and should not be used interchangeably.

The previous Rabi season began without a projected water shortage and with much larger reservoir stocks. Its weak harvest shows that irrigation is not the only influence on output. Reports also describe farmers facing difficulties following changes in wheat policy, though the figures do not isolate the contribution of each factor.

This season's shortage adds another pressure to spending already committed to seed, fertiliser, cultivation and labour. Where usable groundwater is available, farmers may compensate for reduced canal deliveries through pumping. That can preserve crops while increasing production costs.

Where groundwater is saline, inaccessible or too expensive to pump, farmers have fewer alternatives. A similar canal reduction can therefore create different outcomes: higher costs on one farm and reduced output on another. No quantified forecast for the resulting wheat yield loss has been established.

The Pakistan Bureau of Statistics recorded annual food inflation of 8.2% in September 2026. Wheat flour prices rose much faster, by 31.7% in urban markets and 34.3% in rural markets. These increases preceded the coming Rabi harvest and cannot be attributed to a crop that has not yet been produced.

Imports provide a buffer, at a foreign exchange cost

The government initially ordered one million tonnes of wheat imports, including a first phase of 750,000 tonnes intended for November delivery. Later reports describe a reduction to 550,000 tonnes after provincial authorities revised their requirements. These represent different stages of the programme, rather than a single unchanged purchase plan.

The Trading Corporation of Pakistan subsequently finalised 365,000 tonnes at $348.83 per tonne, the lowest price in a September tender whose bids ranged up to $369.95 per tonne. Delivery of the contracted wheat is due by November 20.

Multiplying the contracted quantity by the agreed price gives approximately $127.3 million. At the same unit price, the revised 550,000 tonne programme would cost about $191.9 million. The latter is an illustrative calculation, not evidence that the entire revised quantity has been contracted at that price.

Imports can help cover supply gaps and limit price pressure, but require foreign exchange and expose domestic supply costs to international prices and exchange rate changes. The current purchases respond to the previous harvest and revised requirements. Any additional imports prompted by the new Rabi season remain undecided.

Funding gaps and the next dates to watch

The reported 2026 to 2027 federal development allocation for water projects is Rs 74.9 billion, excluding hydropower generation, within a Rs 1 trillion Public Sector Development Programme. That equals 7.5%. The National Water Policy 2018 set a starting allocation of at least 10% in 2018 to 2019, rising toward 20% by 2030.

On that budget basis, water funding is below even the policy's starting share. Punjab also expects canal water charges of Rs 43.6 billion, while budgeting Rs 8.9 billion for maintenance, about one fifth of expected receipts. These figures point to a gap between collections and maintenance spending, although they do not establish that all receipts must be assigned to maintenance.

The next operational milestones include wheat deliveries and a planned barrage closure. Reports differ on the IRSA approval date, placing it on October 6 or October 7. The principal dates are:

• September 30, 2026: reservoir storage measured 9.343 MAF.
• October 1, 2026: the Rabi allocation season began.
• October 6 or 7, 2026: reports place approval of the 22% shortage forecast on these dates.
• November 20, 2026: contracted delivery deadline for 365,000 tonnes of wheat.
• December 26, 2026, to January 14, 2027: tentative Chashma Barrage closure, coordinated with annual provincial canal closures.
• March 31, 2027: the Rabi allocation season ends.

The closure is described in reports as lasting 21 days, although the stated dates cover 20 calendar dates when counted inclusively. Its exact operating duration needs clarification. IRSA can revise the seasonal shortage estimate, but no specific review date or firm completion date for the delayed Tarbela works has been announced in the reported decisions.

The treaty dispute is separate from the water balance

India placed the 1960 Indus Waters Treaty in abeyance after the April 2025 Pahalgam attack, in which gunmen killed 26 civilians. India suspended treaty obligations, including routine hydrological data sharing and advance notifications. River water has continued to cross the border.

Pakistan's leaders have warned against interference with its allocated waters. Prime Minister Shehbaz Sharif told the UN General Assembly in September that an attempt to divert Indus waters would be treated as an act of war. On October 6, Deputy Prime Minister and Foreign Minister Ishaq Dar warned at the International Water Symposium in Vienna of consequences for regional peace and security.

Those statements concern the international dispute. IRSA's seasonal assessment identifies lower inflows, inadequate storage and domestic infrastructure constraints as contributors to the projected shortage. The assessment does not quantify a loss caused by Indian diversion. Treating the 22% figure as proof that India blocked Pakistan's water would therefore go beyond what the water balance establishes.

Key Points

• IRSA approved a projected 22% Rabi irrigation shortage for Punjab and Sindh, subject to review.
• September 30 storage was 9.343 MAF, about 29.3% below the previous year's 13.214 MAF.
• Lower Kharif inflows and delayed Tarbela works reduced the reserve available for winter crops.
• Balochistan and Khyber Pakhtunkhwa are exempt from the cuts applied to Punjab and Sindh.
• The previous wheat crop was estimated at 27.5 million tonnes, 2.2 million below target.
• Pakistan contracted 365,000 tonnes of wheat for delivery by November 20, at a calculated cost of about $127.3 million.
• The next harvest's output, further import needs and firm Tarbela completion dates are not yet established.