We can’t know the rain in advance, or what İSKİ will decide.
Up to day 30 we give one range. From day 30 to day 90 we show three scenarios, for weather that turns out dry, average or wet. How much water İSKİ brings in from Melen and Yeşilçay is its decision, so that is three assumptions too. Here the scenarios start from today, so the effect of each assumption shows from the start.
The dams in 90 days, under three kinds of weather
These scenarios were produced on 6 November 2026, using data up to 6 November 2026.
Three weather scenarios
For the first 14 days we use ECMWF’s 51 separate weather forecasts. After that we rerun today’s lakes with the weather each past year had on the same dates (26 past years this time). The years are split by that period’s rain into dry, average and wet.
- Dry
- 2000, 2002, 2006, 2007, 2011, 2013, 2015, 2019, 2022
- Average
- 2003, 2004, 2008, 2012, 2017, 2020, 2023, 2024, 2025
- Wet
- 2001, 2005, 2009, 2010, 2014, 2016, 2018, 2021
Scenarios worked out in autumn show too narrow a range at 60 and 90 days. In the hindcast, 80% of outcomes should have landed inside. At day 60 it was 75%, at day 90 68%. What matters most is when the first big rain arrives.
Three transfer assumptions
İSKİ decides how much water to bring to the city from Melen and Yeşilçay, and the weather can’t tell us that. More water seems to come in when the dams run low, but the decision can change at any time. The three assumptions below are approximate daily amounts worked out from the yearly totals on İSKİ’s own page.
- Low transfers
- about 1.07 million m³ a day
- Typical transfers
- about 1.36 million m³ a day
- High transfers
- about 1.53 million m³ a day
We can’t know İSKİ’s future decisions, the state of the source rivers, or breakdowns. Because the daily amounts come from yearly totals, they don’t show the real daily flow.
Consumption: about 3.2 million m³ a day. This is an approximate value read off the charts on İSKİ’s page. It stays that way until a new daily series is published.
What the transfers change
The expected fill level for each scenario and transfer assumption: the range first, the most likely value under it.
| Scenario | Low | Typical | High |
|---|---|---|---|
| Dry | 63–83%73% | 64–84%73% | 65–84%74% |
| Average | 64–85%75% | 65–85%76% | 66–86%76% |
| Wet | 68–91%80% | 69–91%81% | 69–92%81% |
| Scenario | Low | Typical | High |
|---|---|---|---|
| Dry | 68–90%79% | 69–92%80% | 70–93%81% |
| Average | 70–98%86% | 72–99%87% | 73–100%88% |
| Wet | 86–100%94% | 86–100%95% | 87–100%95% |
| Scenario | Low | Typical | High |
|---|---|---|---|
| Dry | 72–100%86% | 74–100%88% | 76–100%89% |
| Average | 84–100%95% | 84–100%96% | 85–100%97% |
| Wet | 86–100%98% | 87–100%98% | 87–100%98% |
How many days the water lasts under these assumptions
This table shows how many days the usable water in the dams would last if the decline worked out in the scenarios carries on. We don’t give a date. We give the result as a number of days, with a range.
None of the three scenarios has the dams losing water over the next 90 days, so a “days of water left” count gives no number today.
If no rain falls at all
This assumes an extreme case: no rain from today, while use and transfers carry on. The record holds no such period. We show it only to see the lower limit.
| Transfers | If the percentages count only usable water | If 10% of capacity can’t be used |
|---|---|---|
| Low transfers | 266 days | 232 days |
| Typical transfers | 295 days | 257 days |
| High transfers | 317 days | 275 days |
We think İSKİ’s percentages already count only usable water. That isn’t confirmed, so we show both.