Case Study Reykjavik

by Remy Mermelstein

Location : Reykjavik, Iceland / Site of Hofdi House

Building Program : UN Commissioned Climate Research Center


This case study looks at the impacts of a 155,000 square foot building placed on the site of the Hofdi House in Reykjavik Iceland. The site is approximately 430’ x 430’ situated along the N shore of the city of Reykjavik with a slope of approximately 15’ from the S to the N (shoreline). The site, as it sits now is empty sans the Hofdi House which sits in the NW corner of the site. The house is approximate 30 x 40 feet and 2-3 stories tall.

To the SW of the site there is a ~275’ tall “skyscraper” which is the among the tallest buildings in Iceland and has caused much controversy because of its height and impacts on the wind along these streets in Iceland.

The buildings on the W/E/S side of the site are all commercial and between 3-5 stories above grade.

In considering the design for the building on the site great thought was put into existing comfort and wind conditions and how a new building would affect those conditions.

An existing conditions simulation was done with Eddy3d, with just the site and surrounding buildings, as well as the Hofdi House, without any intervention to see the conditions present. 24 Wind Directions were simulated, and 3 were picked out as the “worst” conditions:

A UTCI analysis was done as well for the extremes : the coldest avg day in winter, the avg spring day, and the warmest summer day. Weather data was taken from the Reykjavik Airport which sits 2 mi SW of the site in an EPW file from with data for the last 20 yrs.

The script used is available as an Eddy Template, and then slightly modified. We will walk through it here:

24 Wind Directions were chosen, and fed into the Uniform Flow Boundary Condition component, along with the EPW file as stated. The UF BC was chosen given that there is terrain. Later in the final simulation, The Geometry of the building was created into one single mesh, with no windows or details – just the final massing. It is very important that the mesh is closed, and as simple as you can make it without sacrificing too much that is important for the simulation. Remember, at the resolution that this simulation is done, it does not matter if windows/mullions/small indentations/details are present. For this first simulation there was no building, just the site (see below). The Simulation Domain was set up to be 1000m outside radius, and 500m height, about 6x the height of the tallest object in the domain (the skyscraper on the SW corner of the site). For this simulation I did 2,500 iterations.

The results of the simulation as shown above were filtered and 3 were picked out as the “worst” case scenarios:

You can see that in a northerly to notheasterly wind a large wind tunnel is created along the streets running in the similar direction. You can see that the street w/ the tallest building on the corner is much worse than the street a little to the east, and you can see a signature of a vortex forming where even in a N wind, you also get a nasty wind tunnel in the E-W direction along the other main road. Note that this does not happen in the E Street, as that building is quite low (3 floors).

In the 120° direction you can note that the only major issue is again around the skyscraper, with another large vortex from the wind hitting the E corner and creating a vortex.

You can see that our open site allows for some dissolving of this affect in all wind directions. If a tall building were to be built there, or one too close to the street, it could have the potential to create a worse wind tunnel effect along the entire road from E-W south of the site.

The challenge was to create a building that would not exasperate this, but also not completely take away the open site which functions as a park, from the city, nor the view of the mountains from the site (quite beautiful)


The assigned program of the project is a UN Commissioned Climate Research Center. The building had to engage the public, house the private researchers, be home to a world class research center and have a positive impact on the city of Reykjavik.

You can see here the layout of the program, with private and research facilities located on one side, and slowly becoming more public to the right of the building with a central core and vertical circulation. The facility houses several labs (wet and dry), auditoriums, cafeteria spaces, storage spaces, computer labs, offices, administration, and more.

As stated before, the form of the build was derived from 3 ideas : minimizing impact to the existing conditions, giving the park back to the people, and keeping with local Icelandic architecture of low buildings, emulating glaciers and crevices. The crevices would be used for energy generation later in the process with the N wind.

The process for form is detailed below:

With the completion of the form finding a final outdoor wind study was done w/ the massing in place; using the same settings as before for the simulation.

Above you can see for the same wind directions; the vortex from the tall building at the SW corner of the site is still very much present, however, the addition of our building has not made that condition worse. Looking at it in more detail, it is thought that it helped a little bit by blocking some of the upsloping wind from the water making its way S across our site, if anything. You can notice that in a north wind we have a decent wind signature along the crevices of our building. This funneling can be used in the crevices to generate electricity as well.

You can see in this cross section for the 0° wind direction (Wind coming in the direction of the arrows) how the speed is maximized in the crevices.

Pressure variable shows this nicely as well:

In the other cross section for the same wind direction you can see how the wind is pushed upwards at the front of the building by the façade where it is not allowed in the crevices, and then pushed down the slope of the building’s green roof:

You can see that the wind speed decreases at the lowest level closest to the ground as the air is forced higher since the N side of the build is at an elevation as tall as the building on the S side of the Site. This creates a “bubble” of lower wind velocities at ground level on the S side of the site.


By doing this Eddy simulation early in the process, I was able to identify key issues with the site and its surroundings with respect to wind, pressure, and outdoor comfort. Given Iceland’s climate and the influence from the Atlantic, it is not overly cold, and not overly warm. With good material choice and construction, a building could be made to be habitable and comfortable and net positive. A big emphasis was put on the wind condition, as I knew the outdoor comfort was not the # 1 issue. I would not have identified this however without Eddy. Doing subsequent simulations during the form finding (at a low resolution, and low iter count for ease and quickness) helped refine the idea and come to the final simulation and form.