SolVelope is a program which defines and draws solar envelopes. Hoyt Street Yards is a portion of north Portland which is being redeveloped into residential property. This paper examines the application of the solar envelope concept to this area, using the SolVelope program, and proposing a standard protocol applicable to other situations. Weather, orientation, and urban density are considered in determining the appropriate cutoff times to be required of subsequent designs.
As cities develop, access to the sun and the quality of life seem to disappear simultaneously. The fabric of the city is not concerned with solar energy, and all too often is not even concerned with solar access in public spaces. The SolVelope program uses the Solar Envelope concept, which was developed by Ralph Knowles in the mid 1970's, as a possible instrument of public policy, planning and zoning. The envelope is the volume within which the designer must build, in order to avoid cutting sun off from other buildings or public spaces which are off site. (This program was introduced in a previous paper. [1,2] )
There is an area of Portland adjacent to the Williamette River sometimes known as North Downtown which consists of river front, railroad yards, a portion of an old exhibition grounds, and other assorted properties. It is being considered for redevelopment by a consortium of architects, planners, developers, and community activists. (See figure 1.)
It is the intention of the designers to provide a unique character to the development which acknowledges the interface with the river's edge, and seeds the city with environmentally responsive built forms.
There is a portion of the site, the Hoyt Street Yard, which is being considered for medium to high density residential construction. John Caroll of Prendergast and Associates has become interested in developing an optimum set of solar envelopes for this area, so that sunlight access is guaranteed for the occupants of the area, and also to test what sort of group forms result from an area wide application of the solar envelope concept. What form of city fabric is generated by the solar envelope at normal densities?
The question which presents itself is how to employ the solar envelope generating tool. The question is raised in two parts: how should the tool be applied here and is there a general method for applying the tool to other cases? The subset of questions relate to the opportunities presented by the computer version of the tool, and the fact that it is fairly easy to calculate multiple options for cutoff times and edge conditions. Are the correct hours for envelope definition always the same, from one case to another? Can we use asymmetrical timing? What factors should effect, or determine which envelope is chosen? Does such an adjustment process make the envelope concept itself more palatable to developers, and conversely, too amenable to corruption? If the same cutoff hours always apply, there is no room for compromise. If different cutoff times can be examined, the optimum solution can be determined.
There is a second area of examination, as well. What information can be provided to the developer, and subsequently to the designer, which makes the process less cumbersome, and makes the solution itself a better solution? How (in what format) can that information best be transmitted?
Fig. 1. Plan of Riverfront Development Area
In a previous paper, the general form characteristics of various variables were examined. This, plus experience in design classes, indicated that latitudes and daily cutoff times were the dominant factors, and street grid compass rotation, site aspect ratio and street width were the second most important factors. Latitude, in all sites, is an unalterable condition. Latitude, compass rotation, street width and aspect ratio were already established in the Hoyt street case, but this is not true in all cases. The first step in the design protocol should be to consider the implications of the compass orientation of the street grid, block aspect ratio, and street width based on aspect ratio. North-South elongated axes create taller envelopes. But if the blocks are densely packed, the units within the envelope will tend to have either morning or afternoon sun. East-west axis blocks have more low facade (the North side), but more internal units which get even sunlight throughout the day. The North facing units are nearly hopeless. This is in distinct opposition to the typical strategies for low density or single family dwellings.
The second step is to consider daily cutoff times. This may be the most important step. Three criteria are proposed for examination. What are the most critical times of the year, and times of the day, in which solar access is required for direct thermal or energy generation reasons? What are the most critical times of the year and day during which direct solar access should be provided for psychological (and functional?) reasons? What is the solar availability during those times? (It does no good to specify access when no sun is available anyway.)
In Portland, all three of these considerations become significant. The procedure was one of several iterative loops. Weather data was obtained and examined. A site visit observed use patterns, as well as collecting indigenous anecdotal commentary.
Fig. 2. Typical Meteorological Year
The weather data examined was the TMY (National Oceanic and Atmospheric Administrations Typical Meteorological Year) for Portland. The coldest month of the year was December. (See figure 2, which shows one day for each month.) The greatest need for direct solar access was December. Unfortunately, the availability of solar radiation was almost nil. (Again, see figure 2.) A careful examination of the data showed that November through May were months in which solar access was desirable. Fortunately, February through April showed some sunlight availability. The overcast sky tended to break up after 10:00 or 11:00 am, and remained open through 3:00 pm. Thus, solar availability was, indeed, strongly asymmetrical.
This suggested a double break with the traditional envelope. In an attempt to include all of the valuable hours in the critical months, a morning cutoff time from one month and an afternoon cutoff time from a different month could be considered. Including all hours contained in the January 10:00 am to 2:00 pm and the February 11:00 am to 3:00 pm resulted in the inclusion of most hours in which sunlight was useful and available. The resultant envelope is formed by the January 10:00 am planes and the February 3:00 pm planes. (See figures 3, 4 and 5.)
Fig. 3. January 10:00 am to 2:00 pm envelope.
The observed use of open space indicated that the lunch hour was absolutely critical for Portland downtown residents. For example, Pioneer Square is fully occupied at lunch time, and therefor that time is critical. Pioneer Square is a driver for lunchtime traffic, and the resultant shopping. This coincided with anecdotal evidence from the Oregonian natives, who claimed to see sun so rarely that it was a great event to be enjoyed whenever it occurred. Even when there were enclosed malls, etc. the inclusion of an atrium or skylight was recognized as absolutely necessary for successful operation. Thus it was clear that the lunchtime hours had to be included. Fortunately, the weather analysis coincided. In fact, it seems sensible that there would be a congruence between the observed cultural patterns and the weather data.
Fig. 4. February 11:00 am to 3:00 pm envelope.
Fig. 5. Combined January 10:00 am to February 3:00 pm envelope.
The converse should be noted. It might be argued that the December cutoff times should be used, just for the occasional day at which there is direct sun at noon. The trade off, from previous experience, was that the lower sun angles would further clip the envelope volume, which was already impinging on development densities. It remains a judgement question, pitting an ideology we all respond well to against data and marginal return considerations.
The overall height of the envelope is related to the density which might be achieved. The sites which are first developed will be at 50 dwelling units per acre (du/a) but the eventual goal is 75 - 100 du/a. The building massing becomes extremely limited at such densities. This is known from the designs developed at such densities over years of studios run by Prof. Knowles at USC. (See figures 6 and 7.)
Fig. 6. Residential massing under solar envelope at 60 dwelling units per acre.
Fig. 7. Residential massing under solar envelope at 80 dwelling units per acre.
The next question which must be addressed in the protocol deals with choosing how to assemble sites, or literally, how to determine where to draw the edges of individual envelope boundaries. Breaking the site into smaller pieces limits the developer in the flexibility of the design, by reducing the height and volume of buildable space. Conversely, giving too much flexibility to the developer means that individual units within the development may not be protected in the same way that the development itself is protected. This is no small matter in the final quality of the project. Very strong rules about internal zoning must be set by the developer before design proceeds. (See figure 8.)
On the other hand, the design studios have shown that assigning separate solar envelopes to each unit is somewhere between unnecessary and impossible. A tremendous range of solutions is available within the envelope, which still supply solar access to each unit. Obviously, not all rooms in the units receive equal access, nor even all units. But it is quite possible to have all units receiving some sun during the prescribed times at lower densities, or giving all units access to a communal sun space which receives sun during the specified periods (and well beyond, in some cases.)
Fig. 8. Multiple adjacent envelopes.
Finally, if there are functions such as parking structures, strip commercial or industrial processes which do not require, or even reject solar access, the envelope edge should reflect that condition. Standard practice in the design studios has been to raise the line which defines the envelope edge to the ten foot above street level height on the facade in question. Note that this is not the envelope property edge, but rather the commercial property edge (which might be across the street, and may not extend the entire length of the block.) Again, this causes discontinuities in the upper planes of the solar envelope, as was shown in the first paper presented on the subject [1].
In the Hoyt Street project, the envelopes were brought down to ground level in all cases, since the original proposal was exclusively for residential functions on the blocks in question. This might change if the block functions change.
Finally, the envelopes and the appropriate information must be transferred to the developer, who must be able to pass it on to the designer, and perhaps a code enforcement agency, as well. The program screen output can be captured using various utilities, and printed at whatever scale is desired. Tables of the height at every grid point may be printed out for use in checking any point.
However, the most significant improvement in output has recently been developed by another graduate student, Cheng Yu Ho. The program creates a coordinate file which is read by a translator which creates and AutoCAD compatible xx.dxf file. This may be input as a layer in an AutoCAD drawing, and directly superimposed on the design. More importantly, it may be used as a base layer, from which the design itself is developed within the envelope.
The protocol is briefly summarized as follows: