FlowsBook a callSign inCreate free accountSign up
Workflows by industry

How fish farms on the coast watch wind and rain at every site

Strong wind and heavy rain change what a fish or shellfish farm can do tomorrow. How to get a forecast for each pen and bed, and an alert only when it crosses your lines.

Part of the guide Workflow automation

By the xMap team · · 5 min read

A coastal fish farm watches the weather best by setting its own lines, such as the wind above which boats stay in port or the rain above which shellfish beds pause harvesting, and getting a forecast for each site checked against those lines every afternoon. An AI employee can read the forecast at every pen, cage and bed you list, keep only the sites that cross a line, and send the farm managers a short message with what to prepare.

This article is for salmon, sea bass and sea bream farms, oyster and mussel growers, shrimp ponds near the shore, and the advisors who work with them. It covers what the forecast gives for each site, how to choose your own thresholds, a made-up example, mistakes to avoid and the flow that sends the alert.

What a site-by-site forecast gives you

Weather forecast brings a forecast for up to 16 days from national weather services. Given your farm sites as places, it adds to each site the strongest wind expected (wind_max_kmh), the total rain (rain_total_mm), how many days have rain, the first rainy day, the highest and lowest temperature, the UV index and a one-line summary. For a single landing site, a harbour or a drawn stretch of coast you can also ask for the forecast hour by hour, one row per hour, which helps when a boat trip is planned for a precise window.

Two limits are worth knowing. The forecast is for the air above each site: it says nothing about wave height, swell or water temperature, which you still read from your own sensors or a marine service. And hour-by-hour detail is for a point, a city or an area; for each site in a list the forecast is day by day.

Setting thresholds that fit your farm

No two farms share the same limits. A well-moored offshore cage handles wind that would stop a small workboat at an inshore mussel line. Write down the decisions that depend on weather and the number behind each, in the units the forecast uses:

  • Boat work and feeding runs: the wind speed in km/h above which the skipper keeps the boat in. Many farms already have this rule in their safety plan.
  • Harvest of shellfish after rain: the rain in mm over one or two days that usually leads to a pause while runoff clears from the bay.
  • Heat and oxygen: the air temperature that, after a few calm days, tends to come before low oxygen in the pens, so aerators are checked.
  • Net and mooring checks: a wind level that triggers an inspection the morning after a storm.

A made-up example on the Norwegian coast

Invented for illustration. A salmon farm near Bergen runs eight sites: five sea cages in sheltered fjord arms and three more exposed sites further out. The operations lead keeps the sites in a table with each position and a type, sheltered or exposed. The rule for boats is a wind line for exposed sites and a higher one for sheltered sites.

Every afternoon at 15:00 the AI employee reads the next two days at all eight sites. Suppose on a Tuesday the forecast shows strong wind arriving on Thursday at the three exposed sites and nothing worrying inside the fjord. The message to the managers lists only those three, the strongest wind and its day, and xMap AI's line: move Thursday's feeding run to Wednesday, check moorings on Friday morning. The other five sites are not mentioned, because nothing crossed a line.

A shellfish grower on the Galician coast near Vigo could use the same flow with a rain rule: when two days of heavy rain are forecast over the beds, the message reminds the team that harvesting may pause after the runoff and suggests bringing forward the week's orders.

Mistakes to avoid

  • One forecast for the whole farm. An exposed site and a sheltered bay a few kilometres apart can see very different wind. Give each site its own position.
  • Alerts every day. If every afternoon brings a message, people stop reading. Raise the lines until the message comes only when a decision changes.
  • Reading the forecast as sea state. Wind and rain over the site are not wave height. Keep your own marine readings for that.
  • Looking too far ahead. A two- or three-day window is what operations can act on. Sixteen days is useful for planning a harvest week, less so for a boat trip.
  • Forgetting the time zone. Set the schedule in the farm's own time zone so the message arrives before the end of the working day, not at night.

The AI employee that sends the alert

  1. Schedule: every day at 15:00 in the farm's time zone.
  2. My data: your sites, each with a name, a position and its type (sheltered, exposed, shellfish bed).
  3. Weather forecast: Each incoming place, 2 or 3 days, with Wind and Rain (and Temperature in summer).
  4. Filter: A number field, wind_max_kmh at or above your line; add a second Filter or a Branch for rain_total_mm if you also watch rain.
  5. xMap LLM: tell it your rules in plain words and ask for one practical line per site.
  6. A WhatsApp Business step to the managers' group, or Email.

On a calm day nothing is sent. On a day with weather to act on, the managers get a message listing each site that crossed a line, the strongest wind or the rain total, the day it peaks, the weather summary and the one-line advice. The sites stay in your table, so adding a new pen next season is one new row.

Questions

Does the forecast include wave height or water temperature?

No. Weather forecast covers air temperature, rain, wind and UV over each site. Wave height, swell and water temperature need your own sensors or a marine service.

How far ahead can I see?

Up to 16 days. Most farms act on the next two or three days and look further only to plan a harvest week.

Can I have different wind limits for different sites?

Yes. Keep a type or a limit for each site in your table, then use a Branch, or a Calculated field that compares each site's wind with its own limit.

Can the alert go to a WhatsApp group?

Yes, through a WhatsApp Business step from your own account. Email works too.

What about heat and oxygen in summer?

Add Temperature to the forecast and a rule on temp_max. The template for hot days at fish farms does this with feeding and aeration advice.

Templates that do this

Read next

More on workflows by industry · Every article · Templates

Put it to work

An AI employee in xMap Flows does this on live data, on its own timetable. Start from a template or describe it in plain words.