Weather downtime output
Weather is one of the biggest drivers of cost and schedule risk in offshore wind projects, so Shoresim produces several weather outputs. They look similar, but each answers a different question, uses different input data, and is expressed in a different unit. Read each output for the question it answers, and do not compare their values directly.
This article covers the per logistics unit weather downtime figures, the weather downtime causes, and the four weather calendar outputs available for construction cases.
What is weather downtime?
Weather downtime is the time a logistics unit (vessel, helicopter, crane, and so on) could not perform its task because the weather limits for its current process step were exceeded.
All weather limits (for example wave height, wind speed, current) are counted together, and each hour of waiting is counted only once, no matter how many limits were exceeded at the time. Each hour of weather downtime is recorded in the month in which the wait occurred.
How weather downtime is tracked
During a simulation, Shoresim monitors each logistics unit and separates its time into two groups:
- Active time: the hours when the unit is available to work, based on your availability settings.
- Time waiting for weather: the hours when the unit cannot operate because weather conditions exceed your defined limits.
The per unit weather downtime percentage is:
Weather downtime (%) = Time waiting for weather / Active time × 100 Important to know
- Weather downtime ranges from 0% to 100%.
- 0% means no time was lost to weather in that period.
- 100% means there were no usable weather windows at all during that period.
- The calculation only uses time when the unit is available. The following excluded time is never part of any weather downtime or weather calendar calculation: lead time, unavailable, waiting to be scheduled, outside the availability window, outside the work shift, and not spot chartered.
Weather downtime in the dashboard
Weather downtime per logistics unit
This chart shows weather downtime for each vessel or equipment type, in three time intervals:
- Yearly view: accumulated weather downtime for each complete year of the simulation.
- Monthly view (non-aggregated): weather downtime for each individual calendar month as it occurred in the timeline. For example, January 2028 might show 45% for an installation vessel.
- Accumulated monthly view: the same calendar month combined across all years, to show typical seasonal patterns. For example, all January data combined might show 38%.
Weather downtime causes
This breakdown shows which weather parameters (wind speed, wave height, and so on) caused the most downtime. Use it to see where weather limits might be too conservative, or where equipment with greater tolerance would help. When several limits are exceeded at the same time the hour is counted once for weather downtime, but each exceeded parameter is credited in the causes breakdown, so the cause percentages can add up to more than 100%.
In the output report
The per unit weather downtime also appears in the Excel output report (Weather downtime sheet), broken down by logistics unit and time period, with full statistics: average, P10 to P90, minimum, maximum, standard deviation, and confidence interval.
The weather calendar outputs (construction cases)
Construction cases produce four weather calendar outputs. They are related but distinct: each answers a different question, uses different input data, and is expressed in a different unit. Do not compare their values directly.
Key terms used below
| Term | Definition |
|---|---|
| Simulation run | One full simulation of the project timeline with its own weather sequence. A case is made up of many runs. |
| Weather downtime (W) | Hours a vessel waited because the weather limits for its current process step were exceeded. All limits are counted together and each hour is counted once. The hours go to the month in which the wait was recorded. |
| Working time (T) | Hours a vessel spent actively on a process, such as transit, loading, or installation. It excludes weather waiting, waiting on cycle restrictions, and excluded time. |
| Step execution (N) | One completed run-through of a process step, for example one jacking operation. |
| Grouping | All values are reported per vessel, process, and step, by calendar year and month. |
Run-specific weather calendar
Where to find it: Output dashboard, Run specific tab. Can be downloaded as CSV or Excel.
Question it answers: In the selected simulation run, how much of each process's time was lost to weather, month by month?
Which data is used: a single simulation run, the run that matches the percentile you selected (for example P50), and the task group if one is selected. Because the value comes from one run only, there is no statistical spread (no P10 to P90).
How it is calculated: for each vessel, process, step, year, and month:
Weather share (%) = W_step / (T_process + W_step) × 100 - W_step: weather downtime recorded on this step in this month.
- T_process: working time the vessel spent on the whole process in this month. Working time is only recorded per process, so every step in a process uses the same value.
Special rows
- Rollup: the same formula, with W set to the total weather downtime of all steps in the process.
- Average: the same formula, with W and T each added up across all years for that calendar month. The result is one combined percentage for all years together, not an average of the yearly percentages.
Example: a vessel works 75 h on the Loadout process in March and waits 25 h on weather during the Jacking step:
25 / (75 + 25) × 100 = 25% File layout: Vessel; Process; Step; Year; January; February; ... ; December.
Statistical weather calendar
Where to find it: Output report (Excel), Statistical Weather Calendar sheet. Available for construction cases.
Question it answers: How long does one execution of this step typically take, and how much of that time is spent waiting on weather? The answer is based on all simulation runs.
Which data is used: all completed simulation runs. Step durations come from the log of every step execution, with its start time, end time, and duration. Weather downtime is the same as in the other weather calendars.
How it is calculated, stage 1 (within each run): for each vessel, process, step, year, and month:
D = (sum of elapsed duration of all executions of the step) / N W = (sum of weather downtime on the step) / N - D: average elapsed time per execution, in hours, from step start to step end. This includes weather waiting.
- W: average weather downtime per execution, in hours.
- N: number of executions of the step. Each execution counts toward the month in which it started.
- D minus W: the approximate duration of the step without weather.
Note: W is part of D, not added on top of it.
How it is calculated, stage 2 (across runs): the D and W values from each run are combined into statistics: average, P10 to P90, minimum, maximum, standard deviation, and confidence interval.
Time breakdowns: individual monthly (each year and month separately), yearly, accumulated monthly (the same calendar month combined across all years, for example all Marches), and total.
Example: in one run, the Jacking step runs 4 times in March. It takes 40 h in total, including 8 h of weather waiting:
D = 40 / 4 = 10 h W = 8 / 4 = 2 h That run contributes D = 10 h and W = 2 h for March. The sheet then reports the spread of these values across all runs.
Reading the percentiles: P90 of D is the 90th percentile of the average step duration per run. It is not the 90th percentile of individual step executions. Variation between individual executions is averaged out within each run, so the spread is usually narrower than the spread of single task durations.
What is not included: there is no process-level total row. An average per execution only makes sense for a single step.
Typical use: weather-adjusted task durations for project planning tools such as Primavera P6.
Weather downtime calendar
Where to find it: Output report (Excel), Weather Downtime Calendar sheet.
Question it answers: What share of the vessel's total time was lost to weather on each process step? The answer is based on all simulation runs.
How it is calculated, stage 1 (within each run):
Ratio = W_step / (all non-excluded time of the vessel, across all processes, weather waiting included) How it is calculated, stage 2 (across runs): statistics across all runs, as described for the statistical weather calendar (average, P10 to P90, minimum, maximum, standard deviation, confidence interval).
Notes
- The ratio is relative to the vessel's total time, not the time spent on this process. A process that rarely occurs therefore shows a small percentage, even if weather hit it hard.
- A process-level row shows the total for all steps in the process.
Typical use: understanding how much of a vessel's overall time is lost to weather, and on which activities.
Net weather calendar
Question it answers: How much of each process's time is lost to weather? The answer is based on all simulation runs.
How it is calculated, stage 1 (within each run):
Ratio = W_step / (T_process + W_step) This is the same formula as the run-specific weather calendar. The difference is that the net weather calendar reports statistics across all runs, while the run-specific calendar shows a single run.
How it is calculated, stage 2 (across runs): statistics across all runs.
- A process-level row shows the total for all steps in the process.
Typical use: process-level weather exposure, expressed as a statistical distribution.
Comparing the weather calendars
| Run-specific calendar | Statistical weather calendar | Weather downtime calendar | Net weather calendar | |
|---|---|---|---|---|
| Runs used | 1 (selected percentile) | All | All | All |
| Unit | % | Hours per execution | Ratio | Ratio |
| Numerator | Step weather downtime | Duration (D) and weather downtime (W) | Step weather downtime | Step weather downtime |
| Divided by | Process working time + step weather downtime | Number of executions | All vessel time | Process working time + step weather downtime |
| Statistical spread | No | Yes | Yes | Yes |
| Process total row | Yes (Rollup) | No | Yes | Yes |
| Typical use | Explaining one scenario | Task durations for planning (for example P6) | Vessel-level weather exposure | Process-level weather exposure |
Important: These outputs are measured in different units. The statistical weather calendar is in hours per execution, while the others are percentages or ratios of time. Do not compare their values directly.
Understanding statistical values (percentiles)
Each construction case runs many times with different weather sequences, so the results form a distribution. The percentile values describe that range:
- Lower percentiles (P10 to P30): more optimistic outcomes.
- Middle percentiles (P40 to P60): most likely outcomes. P50 is the median.
- Higher percentiles (P70 to P90): more conservative outcomes.
For planning, many users focus on P50 for a balanced view and P75 to P90 for conservative, risk-informed planning.
Tips for using weather downtime data
- Compare logistics units to see which vessels lose the most time to weather.
- Use the accumulated monthly view to find the best and worst months for operations.
- Review weather causes to see which parameters are the binding constraints.
- Use the weather calendars to find which specific operations are most weather-sensitive, and choose the one whose question and unit match your need.
- Use P75 or P90 when you need to build weather contingency into a plan.
Frequently asked questions
Why don't the run-specific calendar and the net weather calendar show the same numbers? They use the same formula. The run-specific calendar shows one run, while the net weather calendar shows statistics across all runs. The run selected for a percentile (for example P50) is not guaranteed to have the P50 value for every process and month.
Can I compare the statistical weather calendar to the other calendars? No. The statistical weather calendar is measured in hours per execution, while the others are percentages or ratios of time.
Why do months in the statistical weather calendar sometimes look inconsistent? A step execution counts toward the month it started, while weather waiting counts toward the month it occurred. A step that starts at the end of March and waits on weather into April adds its waiting time to April. Yearly and total values are not affected.
Why are some months empty? A value only appears where the vessel actually worked on that process or step in that month.
Why do I see different weather values in different outputs? Each output answers a different question. The per unit weather downtime is a share of a vessel's available time. The weather downtime calendar is a share of a vessel's total time, per step. The net weather calendar and the run-specific calendar are a share of a single process's time. The statistical weather calendar is a duration in hours, not a share. All are useful for different planning purposes.
My weather downtime seems high. Is that normal? Weather downtime varies greatly by location, season, and operation type. North Sea operations in winter typically have much higher weather downtime than summer operations in calmer seas. Compare your results against benchmarks for your location.
How do the weather limits I set affect weather downtime? Tighter limits (lower maximum wind speed, lower wave heights, and so on) increase weather downtime. If downtime looks very high, check whether your limits are realistic for your equipment and operations.
Why is the weather calendar broken down by process and step? Different operations have different weather sensitivities and weather window requirements. Breaking the data down by process and step shows which specific activities are weather bottlenecks and may need special attention or different limits.
Need more help? Contact our support team for help interpreting your weather results and optimising your simulation setup.