Reading a climate before you design in it

A site brief will give you an average temperature. It is almost always the air temperature, measured in a shaded ventilated box at two metres, and it is not what anybody standing on the site experiences. Wind strips heat off a body. Sun adds it back. Humidity decides whether sweating works.

Run the two files in this folder and the size of that difference stops being an abstraction. Cork's air averages 10.47 C across the year. What a body feels averages 2.82 C. The gap is 7.65 degrees, which is larger than the 6.8-degree difference between Cork's air and Rome's.

So the first move on any site is to find the gap, and the second is to find out what makes it. In Cork it is wind, and a design that shelters without trapping is worth more than a design that insulates.

UTCI answers one question: given this air temperature, this mean radiant temperature, this wind speed and this humidity, what air temperature in a reference condition would feel the same. So the output is in degrees Celsius, and it is a felt temperature rather than a measured one.

The index carries published stress categories, and these are its own terms. Do not rename them to fit a legend.

An EPW holds 8760 rows, one an hour for a year, with temperature, humidity, wind, radiation and sky conditions on each. The year is not a real year. A TMYx is assembled by picking, for each month, the most typical instance of that month from a stated span of real years, then stitching the twelve together.

The files here are TMYx 2011-2025, so they describe the fifteen years to 2025. That has two consequences you carry into every figure you report. The file is a record of the recent past. And the year span is part of the citation, because a Cork file assembled from 2004-2018 gives different answers.

Ladybug Tools has a Python core, and the Grasshopper components are a canvas over it. Installing that core gives you the same engine without Rhino, which matters twice: a learner without a licence can still do the analysis, and everybody gets a reference to check the canvas against.

python -m venv venv
source venv/bin/activate      # macOS or Linux
venv\Scripts\activate         # Windows

pip install -r requirements.txt

Cork and Rome are already in weather/, so there is nothing to download to start.

First the script, so you have a number before you have an opinion.

python utci_from_epw.py weather/IRL_ST_Cork.AP.039550_TMYx.2011-2025.epw

You are looking for a UTCI annual mean of 2.82 C and 30 hours at no thermal stress. Those two numbers are your target for the rest of the session.

Then open Grasshopper and build definition 3 from the definitions sheet: import the EPW, derive outdoor mean radiant temperature, wire air temperature, humidity, wind and that MRT into LB UTCI Comfort.

Two settings decide the answer and both belong in your dossier. Wind is recorded at 10 m in an EPW and the index is defined at 1.1 m, so the component reduces it. And mean radiant temperature is not in the file at all: it is derived, assuming a person standing in the open with nothing shading them. Change either and you have answered a different question, so say which one you answered.

1. Read the weather file. The one everything else starts from. File path into LB Import EPW, dry bulb into a monthly chart.

2. Sun path and radiation. Location into LB SunPath, direct and diffuse radiation into LB Cumulative Sky Matrix and a sky dome. Cork's path is visibly flatter than Rome's, and the difference is ten degrees of latitude.

3. UTCI. The one the module turns on, described above.

4. Honeybee daylight. Where your own geometry enters: a room as a closed brep, apertures by ratio, then annual daylight through Radiance. Start with one room and a coarse grid, because Radiance is slow. A fine grid over a whole model will run for a long time.

No .gh files ship with this module. A Grasshopper definition is a binary file written by Rhino, and nothing in this repository can author one. A file that claims to be a working definition and turns out to be empty is worse than no file, so the sheet gives the component sequence and you assemble it once. Assembling it is the better exercise: a definition handed over as a black box teaches you to press a button.

UTCI as computed here describes a person standing outdoors in the open, in the wind recorded at an airport weather station, with the sky above them unobstructed. Almost no interesting place in a city is like that.

A courtyard is sheltered, so its wind is a fraction of the recorded figure and its UTCI is much higher. A north-facing lane never gets the sun the file reports. A street canyon radiates stored heat back at night. None of that is in these numbers, and reporting a station UTCI as though it described your site is the error this section exists to stop.

What the station figure is good for is the climate the site sits in, the size of the gap between air and experience, and what makes that gap. Those hold across a city. The moment you want a figure for one place, you need the microclimate, and that is what infrared.city does next door.

Read the minimum row twice. Cork and Rome bottom out within a degree and a half of each other, which is not what anybody expects. The two climates differ in how much of the year sits in the comfortable band, and hardly at all in how cold it can get on the worst hour.

The maximum row is where they part: Rome reaches 43.25 C on the index, which is very strong heat stress, and Cork never leaves moderate. A design move that is obviously right in one column can be obviously wrong in the other, and the fortnight in Rome is when that stops being a table and becomes something the cohort stands in.

infrared.city takes the next step: a street that already exists, read as UTCI, wind and solar grids, with the question of whether added canopy actually cools it. This module gives the climate, that one gives the microclimate.

The academy holds the Revit and Dynamo side: Revit fundamentals, two family guides, the Dynamo workflow, and the Jitterbug parametric family automation that runs the synergetics geometry into a Revit family. Where this module reads a climate, that one builds the thing that sits in it.

Emergy Explorer on ENGAGE, for Odum's energy accounting, and AECO2 BIM Canvas for actors, stocks and flows.

Into your own dossier. Every project in the module carries a climate reading. Name the weather file with its year span, say whether wind and mean radiant temperature were included, and report the gap as well as the mean. A UTCI figure without those is not checkable by anybody, including you in six weeks.