Cholera Treatment Center in Haiti

July 9, 2012 / no comments

Recently we were asked by MASS Design Group to do a daylighting study for their new Cholera Treatment Center (CTC) in Port-au-Prince Haiti. This is the second time that we have had the opportunity to work with them. Since the earthquake on January 12, 2010, Haiti has suffered enormous economic, structural and environmental distress. Cholera, which previously did not exist in Haiti, broke out shortly after the earthquake and according to the World Health Organization, as of November 30, 2011, there were 515,699 reported cases of cholera and 6,942 deaths.i The new CTC will be a year-round center for the care and treatment of patients suffering from the illness. This project presents an unusual opportunity to engage lighting on such fundamental levels, and to think about basic human needs as they relate to light. It’s even more unusual to be facing these issues in a country like Haiti. As designers in the United States, our typical projects include commercial, residential and institutional projects in major developed cities. While many basic principles of good lighting remain universal, working on a project like this exposes us to significant cultural, financial and climatic differences.

What we learned

MASS Design Group is a non-profit architecture and design firm dedicated to helping improve the health and overall well-being of communities through design. Their most well-known and highly-praised projects include the Butaro Hospital and the Girubuntu School, both in Rwanda.

During our design conversations, members of the MASS team shared with us examples of the kinds of design and technology challenges that they have encountered in their work. One team member, Elizabeth Timme, shared a few experiences from Butaro Hospital. The hospital is tailored for the treatment of tuberculosis, so enormous fans are suspended from the ceiling and UV lights shine from the walls. The fans keep air moving (a critical element in tuberculosis treatment), and the UV lights kill airborne bacteria. Patients however, worried that the fans were “stealing their air” and that the lights were “burning their skin”. To some extent both comments are true, but negotiating the relationship between the function of the space and the occupant use and comfort highlights a fundamental mission of design.

For decades Western hospitals were built to enhance efficiency and hygiene, and only recently has there been more evidence that daylight, views and good design are equally critical factors to providing the best health care. In fact, Timme is now starting her own firm, Más, in order to bring a similar design approach as used at MASS to the issues facing the American health care system.

MASS has encountered other issues that have to do with climate and resources. In one case, they told us that in Rwanda, surface brightness was a serious design consideration. The sun at that latitude can cause surfaces to be so bright that they create visual discomfort. An article by Martin Schwartz about Louis Kahn’s proposal for a U.S. Consulate in Luanda illustrates this issue. In it Schwartz quotes Kahn:

“I …noticed that when people worked in the sun-and many of them did-the native population …usually faced the wall and not the open country or the open street. Indoors, they would turn their chair toward the wall and do whatever they were doing by getting the light indirectly from the wall.” ii

Not only do these types of realizations help us to understand better how to work in under-served countries, but they help to inform our approach to design for our typical projects. Working on a cholera treatment center can help us to recognize and consider factors in our day-to-day work that we may not have previously considered to be important or necessary.

What we did

The center is 7,700 square feet and consists of a general ward, an intensive care ward, as well as an administration office. The roof consists of 15 roof modules, each 23′ x 23′ oriented in different directions. Four central modules serve as the top of a cistern designed to collect water, while the remaining 11 are pitched, both to allow for light and ventilation as well as to direct water towards the cisterns. The building is primarily open on all sides, although there are fixed screens on the North and East sides and sliding screens hung on the West side.

Since construction had already begun when we joined the team, the project manager, David Saladik, emphasized that we needed to quickly and efficiently produce a series of clear and useful daylighting studies. There was no possibility for bells or whistles. We had to make simple and effective recommendations that could be executed within the time and resources available. It was important to set our criteria immediately, and use tools that would help us achieve our goals. For all of the simulations we used the daylighting program DIVA-for-Rhino.

Our analysis focused on three key issues: sufficient light levels, glare, and heat gain. To test for adequate light levels, we ran a daylight autonomy test using a horizontal calculation grid at the height of the beds, and vertical grids at the headboards. We used a threshold of 500 lux because light levels need to be relatively high, since nurses and doctors will be conducting procedures and treating patients at their bedside. We were pleased to see that the daylight autonomy results for the existing design showed that we can expect 500 lux over the majority of the ward space for 50% of the year.

1_Blog CTC DA 500 plan.jpg

Daylight Autonomy calculation (500 lux threshold)

Having established a general sense for annual daylight levels, we wanted to look more closely at the lighting conditions during specific points in time when we could estimate there might be problems. The East side is most vulnerable to overlighting and heat gain because there are no structures directly to the East of the center. We wanted to test a variety of material options for a proposed set of panels, which would serve as shading. We chose three materials to test: opaque panel with 50% reflectance, a translucent panel with 40% transmission, and 40% open screen.

2_Blog CTC Screen Analysis.jpg

Evalglare Glare Analysis

Summer Solstice 9am Glare Analysis 40% Open Screen (top right)

Winter Solstice 9am Glare Analysis of (3) different panel options (bottom)

Our results show that only the opaque panels will have a significant effect on minimizing the overlighting on the East-side ward. The translucent and 40% open screen still allow a significant amount of light and, by extension, heat into the building. In addition, when we ran the glare tests for the three material options, the 40% open screen performs the worst in terms of glare for the patients in East-facing beds. This only gets worse during the rest of the year when the sun is lower in the sky; on the Winter Solstice at 9am, the Evalglare Daylight Glare. iii Probability result showed there would be intolerable glare . Since the preferred solution is to use the screens, we recommended modifying the open percentage of the screen to be less than 40%.

3_CTC Roof Analysis.jpg

Glare experienced from hospital bed on Summer Solstice at Noon (bottom left)

Roof module and assembly (bottom right)

Cholera outbreaks in Haiti are the most frequent during the rainy season, which happens in the summer. Knowing that, we were particularly interested in the summer conditions since those are the times the wards will likely be most full. In particular, we looked for problems with glare and overlighting. The orientation of the roof modules on the East side allowed, rather than prevented, direct early morning and noontime sun to stream into the building. We confirmed this by running several glare simulations, and made the recommendation that two of the roof modules be rotated 90 degrees in either direction to eliminate the problem.

4_Blog CTC Trough Analysis.jpg

Illuminance Calculations Summer Solstice 12pm

Existing Glazing Design (left). Glazing replaced with 80% Opaque Panels (right)

The last issue we wanted to tackle was the glazed portions of the roof that would direct water into the cisterns. Those sections would be glazed in order to let light into the center of the building, and be directly above a small, planted garden. We ran several tests on the Summer Solstice, which showed that there would be dramatic overlighting, glare and potentially great heat gain. While we understood the design intent to create a small lush area in the middle of the ward, we suggested some simple louvers to mitigate the quantities of incoming daylight.

While absent of high-technology or complicated details, our solutions answered the key daylighting questions and will contribute to providing the patients and staff with the most functional and comfortable space possible. Even with the demanding constraints of a project like this, in the end, we’ve shown that with targeted thinking and the right tools, we can come to useful conclusions and make effective recommendations, which is definitely a lesson we can bring back to our everyday work.


i World Health Organization, “Health Cluster Bulletin: Cholera and Post-Earthquake Response in Haiti”, 21 December 2011.

ii Louis I. Kahn Writings, Lectures, Interviews, page 123, quoted by Martin Schwartz in “Louis I. Kahn: Finding Daylight in Luanda”, February 7, 2011.

iii Evalglare calculates glare according to several factors such as brightness and size of glare source, and gives a result called the Daylight Glare Probability, which was calibrated by user-assessments. The categories of glare perception from least severe to most are: Imperceptible, Perceptible, Disturbing and Intolerable. For more information see: Evalglare.

Image Credits: Kera Lagios/Lam Partners

Caveat Metrics

May 11, 2011 / no comments


Daylighting metrics are methods for measuring the quantities of daylight in a space during a period of time. More and more, metrics are becoming the dominant means by which daylighting in a space is evaluated. With the imminent adoption of the International Green Construction Code and other codes mandating daylighting, the use of metrics will become even more integrated into the daylighting evaluation of buildings. While evolving analysis tools provide new and exciting capabilities, they also present new challenges to the designer or consultant.


Metrics have the inherent benefit of providing better information on the performance of a space than traditional rule-of-thumb methods. They are fast, adaptable, and instill confidence in the client, and the flexibility of digital modeling allows many variations of a design to be tested quickly at early stages of the design process. Unlike rules-of-thumb, metrics are more easily capable of evaluating non-orthogonal spaces, and they are becoming more accessible as more and more software provides daylighting analysis tools. And if that were not enough, increasingly, clients demand to see statistics and false-color grids in order to be convinced that their building will perform well, achieve credits, or meet codes.

But, like all things, metrics have downsides: metrics can be deceptively convenient. It seems as if it should be relatively easy to just build or import an architectural computer model into a simulation program and run the metric, but this is not so. Each software has its own rules for producing correct output. These include ways in which geometry should be modeled, whether or not to include a ground plane and how to define materials. Different lighting simulation engines have different ray-tracing methods (e.g. backwards versus forwards), and different simulation settings. The same basic variable likely has a completely different name from one program to another, and of course, the software interfaces are different – certain programs allow control over lighting variables, while other programs keep the user from accessing or modifying those variables.

Conversely, one benefit to the increased focus on daylighting metrics is their increasing accessibility. Plug-ins like DIVA-for-Rhino and the su2rad script allow widely used softwares like Rhino and Sketchup to interface with Radiance, the premiere calculation engine. While this overall accessibility is positive because it allows daylighting analysis to be employed more freely, making it more of a player in design decisions, it also makes education about the proper use of those metrics much more important.

The first step in understanding metrics is to know what metrics currently exist and what information they can provide. A useful guide to daylighting terminology was provided by Kevin Van Den Wymelenberg in an Architectural Lighting article in 2008, where he defines several of the daylighting metrics currently most in use today: Illuminance, Daylight Factor (DF), Daylight Autonomy (DA), Continuous Daylight Autonomy (CDA), and Useful Daylight Illuminance (UDI).

As a quick overview, the main distinction between various metrics is between the so-called “point-in-time” (Illuminance, Luminance) and annual, climate-based calculations (DA, CDA, UDI). Point-in-time calculations measure light levels at a specific date and time, under a specific sky condition. These calculations are more intuitive because they mimic how we experience the world: we see the light levels change from one moment to another. Annual or climate-based calculations, on the other hand, use weather data to simulate lighting levels over the length of an entire year. As such, they are more comprehensive than point-in-time metrics, but are also a more abstract, less intuitive way of measuring lighting. While they provide a more comprehensive performance evaluation, they may not show as clearly why one scheme performs better than another. Daylight factor, which originated in the cloudy climate of Britain, is neither point-in-time nor annual, as it uses an evenly illuminated (overcast) sky condition to measure interior-to-exterior light ratios.

Once designers have some idea of which type of calculation to use, they are faced with the issue of whether or not they can use it. Currently, the majority of lighting calculation software provides only illuminance and luminance calculations on a point-in-time level (for example, a clear day on September 21st at 9:00 AM). In general, there is a movement towards using annual, climate-based calculations rather than point-in-time, but the critical issue is that most commonly used daylighting programs do not support climate-based metrics. At present, 3dsMax and AGi32 only calculate illuminance and luminance (point-in-time). Daysim is the only widely used lighting engine which can perform the annual calculations.

The given metric may not really deliver answers to the questions at hand. From an architect and owner’s perspective, there are usually several critical questions posed to the consultant about daylighting. The first two are: how often will we be able to dim or turn off the electric light, and how will daylighting affect thermal performance? Currently, there is no good metric to directly answer those questions. Christoph Reinhart and Jan Wienold have developed one metric, called Daylight Availability, which perhaps comes the closest. In their paper “The Daylighting Dashboard – A Simulation-Based Design Analysis for Daylit Spaces,” they document the metric. It combines DA (Daylight Autonomy) and UDI (Useful Daylight Illuminance), and shows, in one false-color grid, the assessment of areas that are likely to be overlit (requiring shading), well lit by daylight alone, or partially daylit (requiring supplemental electric light). It is possible that this metric, or one like it, could fill the void.

The final part in the daylighting metrics process is the output. Once a metric has been chosen and run, the programs produce either a rendered image, a false-color image, or a grid of numbers as a result. The job of the daylight analyst is done, right? Of course not. This step can be the most challenging of all. Expressing daylighting analysis results in an intelligible way, and presenting them to a client can be difficult. There is no formula for the best way to do it, and it often comes down to what the particular situation requires. The fact is that it is difficult to synthesize in a single image the variability of lighting conditions over the day and year, and when multiple design options like shading devices, materials, or orientations are added, the complexity expands proportionally. Given this, there is a tendency to become metric-happy and produce copious studies for different times and under different conditions; this often overwhelms the client who, unfamiliar with the format, may barely understand a single false-color grid, let alone a set. Even for sophisticated daylighting designers, the useful conclusions may be hidden in the sheer mass of output.


Outputs produced with DIVA-for-Rhino

There is no single metric which can answer all questions; each provides only part of the story. Annual calculations provide information about lighting levels, but not about glare, thermal costs, or aesthetics. One idea beginning to gain acceptance as a solution is the concept of a “dashboard”. Dashboards, as laid out by Reinhart and Wienold, are meant to show summary results of many metrics in a single side-by-side view, although, it should be noted, that synthesis is still left to the consultant.


Reinhart and Wienold, “Daylighting Dashboard” concept image

Lastly, the architect’s and owner’s question, “What will it look like?” still prevails. False-color grids and numbers don’t read as quickly as does an image, and after all, a large part of the value of daylighting design is improving the visual quality of the space. Images may contain the least amount of hard data, but they tend to go the furthest in illustrating daylighting concepts to clients.

As we enter this new phase of daylighting analysis, it is important to know the strengths and shortcomings of each metric and to be informed as to how to properly use them. The increased predominance of the computer does not change the fact that it is the designer who must know how to use the tools, how to understand the results, and how to effectively communicate the results to team members and clients.

5_DIVAforRhino Visualization_KLagios.jpg

Radiance Visualization using DIVA-for-Rhino

Images credit: Kera Lagios(1-3,5), Christoph Reinhart and Jan Wienold (4)