Winter Greenhouse

PROJECT BACKGROUND
I own and operate a small farm of about 70 acres in White, SD. Most of it is dedicated to pasture, while I farm organically about 10 acres. I have kept goats and layers, but am presently transitioning to more intensive farming. The winter greenhouse project began soon after I retired and moved to SD in 2006. It was my background as a physicist that suggested the SARE Winter Greenhouse project.

When I moved onto this farm in 2006, I had always intended to farm it organically, employing sustainable agricultural techniques. With this in mind, I have only used organically approved fertilizers and herbicides (mostly ineffective) on the farm. Since taking 10 acres out of CRP into farm land, I have had considerable problems with thistles. To cope with this problem I have employed almost exclusively mechanical strategies (e.g., mowing and discing) in conjunction with competitor planting (e.g., rye, buckwheat, and alfalfa). Additionally, I have employed rotational grazing techniques. The layers that I have kept have all been free-range pastured. The very conception for the Winter Greenhouse project is motivated by sustainable notions. It is for this reason its design and strategies have been exclusively focused on passive applications.

PROJECT DESCRIPTION

Goals: The initial and enduring goals of the Winter Greenhouse Project have been to develop low cost passive strategies for winter greenhouse production. To date, the primary focus has been upon establishing those ambient beneficial conditions appropriate for winter greenhouse production. With this in mind, attention has been primarily upon soil and air temperatures under various strategies. Additionally, it is the goal of the project to share both any information gleaned and the instruments employed.

PROCESS
The first step was designing and building my winter greenhouse. The design of this greenhouse was influenced heavily by the work of the University of Missouri. At the time, I knew of no one in the area who had or was designing a winter greenhouse. (Chuck Waibel and Carol Ford in Milan, MN apparently had just completed the construction of their winter greenhouse at this time.) I knew that environmental conditions in Missouri were considerably less challenging than those in South Dakota. Having some experience with small building construction, my major concern was for the design of the large solar window. I wrote a number of computer programs to determine the angle of the solar window. Having made that determination, my next concern was for the material from which to construct the solar window. Other winter greenhouses (e.g., Missouri, University of Manitoba, and Chinese solar greenhouses) use 6 mil plastic, usually with a double layered to increase R value. I experimented with a number of materials before meeting Carol Ford and Chuck Waibel. I found that 6 mil plastic and like materials could not handle the South Dakota winters and winds. I think it was in 2007 that I met Chuck and Carol. They were using 10 mm double walled polycarbonate for their solar window. This is a rather expensive alternative. But for small winter greenhouses it might be an acceptable alternative. In the four years that I've been using 8 mm double walled polycarbonate I've had no trouble with the solar window, and I know that Carol and Chuck's experience is similar. Given the durability, and life time of the double walled polycarbonate and the time and cost of replacement of cheaper materials, double walled polycarbonate ought to be seriously considered as a cost-effective alternative. It was also from Chuck that I considered the value of insulating the perimeter of the greenhouse down to about 4 feet, something I completed about two years later.

Over the years, through both contacts with others, my own experience, and computations, there are number of elements of the winter greenhouse design that I would either do differently or consider more carefully. Included in these considerations would be the likes of (1) the solar window angle for the winter greenhouse and its relationship to greenhouse volume; (2) whether or not to have a partial south wall; (3) the use of raised planters; (4) thermal storage in north wall vs water barrels; (5) insulation of the perimeter of the greenhouse, a cost-effective analysis; (6) consider building the greenhouse into a hillside; (7) consider whether to have east and west solar windows or not; (8) consider carefully the location of the greenhouse for winter solar irradiance; (9) consider the placement of heating elements in the soil and whether they would be heated by air or water.

After construction of the winter greenhouse, I next considered the kinds of strategies I wanted to investigate. Over the course of greater than two years of the project I have principally considered (1) the use of water-filled barrels as thermal storage; (2) utility of insulating the perimeter of the greenhouse; (3) effectiveness of a solar curtain; (4) efficacy of hanging planters vs raised planters; (5) effectiveness of various soil coverings; and (6) the utility of a water thermosiphon for heating soil.

With these and other strategies, how can one access their effectiveness? Since economics ought not be neglected, it must be possible to make some kind of cost-benefit analysis. But to do this more than a qualitative analysis is required, rather some kind of quantitative one. It is possible to imagine such an analysis being carried out over the course of several seasons of experience. However, a more direct, precise, and rapid analysis might be performed with the aid of a detailed instrumentation of the greenhouse environment. With this in mind, what kind of instruments would be required? We would want to measure soil temperatures, internal and external air temperatures, water temperatures, and solar irradiance. More than this, one would require some means of logging the data automatically since otherwise data collection is likely to be too sparse (if not overly inconvenient). With this initial assessment, it was a SARE grant that funded the bulk of these initial procurements. Over the years, I've discovered that I need many more temperature sensors than the four I initially purchased. This is because the three-dimensional variation in greenhouse temperatures of soil, air, and water is considerable. This would likely not be true for a much larger greenhouse.

Classical statistical analysis would assess the efficacy of any strategy by establishing a control and by comparing the differences in some measurable and significant entity or entities between the control and the treated samples. We might, for example, examine soil and water temperatures. We would then compute the sample means and variances for the two cases. Having done that, we need to be able to assess with some probability the difference between the two actual means, and not simply between the two sample means, which are themselves random variables. In order to make progress we would usually presume that the true variances (as opposed to the sample variances) are equal. While this is likely not exactly true, it likely that they are close to each other. With this assumption it is now possible to employ what is called the Student T probability distribution. This probability distribution is derived on the assumption that the random variables (soil and water temperatures) are normally distributed. It is then possible to form a random variable, using the sample means and variances, that is independent of the actual variances (presumed equal), but dependent upon the true and actual means. Having gotten this far, one then arbitrarily establishes a confidence limit. The confidence limit establishes the probability of a false negative that you are prepared to accept. A false negative is what happens when one uses some criterion to decide when the probability of a certain event is so low that one judges it could not happen yet it does happen. The confidence limit is established by choosing the total probability of a false negative that one is willing to bear. Once this confidence limit is established, one can determine, using the Student T distribution, for the two sets of data (e.g., soil temperatures with and without the solar canopy) a confidence limit for the difference between the actual means. The result will be a certain confidence that the difference between the two actual means is between two values.

What is outlined above is the normal statistical approach to evaluating statistically the efficacy of a certain difference from the control. There are, however, problems with this approach. Suppose we want to evaluate the efficacy of a solar curtain used at night. In this case, the control would be the greenhouse absent the solar curtain, and the treated sample that of using the solar curtain at night. The problem is that there are a number of other factors that significantly effect the soil and water temperatures, including the solar irradiance, external temperature, and wind velocity. It may even be that cloud cover affects the results since more cloud cover will increase the diffuse component of the solar irradiance and that might influence the amount of radiation that gets into the greenhouse. All of these other independent variables have a significant effect upon the soil and water temperature history. How can we be certain, then, that we are actually measuring the affect of the solar curtain and not some other variable? The traditional means of dealing with this problem is termed randomization. By including in both datasets a wide range of these uncontrolled variables, we hope that their effect will average out. Ideally, we would like to have the same kind of variability in the uncontrolled variables in both datasets. Perhaps the best way to ensure this is by having large datasets. Taking datasets over the course of a year, or even longer, would provide some confidence that the rich effects of these diverse uncontrolled variables have been sufficiently averaged out.

The problem in being able to make such an evaluation for the solar curtain is that there is only one greenhouse. As a result, simultaneous measurements cannot be made of the control and the treated samples. Even in the case of trying to assess the efficacy of soil coverings physical location in the greenhouse can produce significantly different thermal histories independent of the soil coverings. One can try to try to overcome this bias by exchanging the treated and control locations and averaging the two results. In the end what this generally entails, is that one can only measure effects that are significantly above the noise generated by the variation in location and time associated with the greenhouse variables.

The last step in this process is the evaluation of various greenhouse computational models. Such models can be simple and parametrized to fit a particular dataset. Others would be more detailed starting from first principles. In any case, comparison between datasets and numerical predictions is required.

PEOPLE
Chuck Waibel provided some assistance early on by sharing the design of his winter greenhouse in Milan, MN. Subsequently, I did follow his lead in using twin walled polycarbonate for the solar window and in burying R20 foam insulation 4 feet about the perimeter of the greenhouse. Chuck's winter greenhouse takes air from the top of the greenhouse, where the temperatures are the highest, and forces it to circulate through a rock bed in the soil. For reasons that were probably mistaken, I built my winter greenhouse upon an extant cement slab. For that reason I could not retrofit a similar soil heating system. I did later implement a passive system using water, which I believe is superior to Chuck's system. I could be wrong about this, but in any case, the thermosiphon I employ is passive, whereas Chuck's requires an electric fan.

Carol Ford is a valuable local resource for the kinds of greens that will grow in the upper Midwest in her winter greenhouse. Although in recent years I have focused more on the evaluation of strategies independent of plant production, ultimately I would want to test those strategies in the context of winter production.
Greg Mishna of South Dakota State University Engineering Department was helpful early on in helping me to develop simple greenhouse heat transfer models.

Dr. Qiang Zhang of the University of Manitoba Biosystems Department has helped both in print and with some personal communications. His journal article was the first that I examined modeling a winter greenhouse (Zhang Q., et al, Winter performance of a solar energy greenhouse in southern Manitoba, Can Biosystems Engr, v 48, p 5.1-5.8, 2006). I have also profited from some private communication relative to various solar curtain strategies, including having the solar curtain on the inside or outside of the solar window, and the use of argon filled bags to raise the R value of the solar window.

Lawrence Berkeley Laboratory Staff have been helpful with regards to their window modeling codes. These codes are freely distributed. I found them useful when I was attempting to both design a thermal solar curtain and in theoretically determining the thermal properties of the solar curtain I had constructed.

Elmer Staff: Elmer is a 3D finite element code that is freely distributed. I spent some time trying to create a 3D model of my greenhouse. They were very helpful in this regard. I have not completed this task as I ran into some difficulties with regard to modeling convection. But I do plan to have another go at it at some time in the future.

RESULTS
I spent considerable time attempting to confirm the existence of a significant reduction in heat loss with the solar curtain. My greenhouse heat loss models predicted greater than a 10 percent overall reduction in heat loss, and a 40 percent reduction in heat loss through the solar window. Some papers that I read (e.g., Bailey, B.J., The reduction of thermal radiation in glasshouses by thermal screens, J Agr Eng Res, v 26, p 215-224, 1981) confirmed the significant gain achieved with a solar curtain. Zhang at the University of Manitoba and all the Chinese Solar Houses used solar curtains. I spent a good deal of time trying different low cost strategies. Ultimately the one I settled on was a single layer of Reflectix that was raised by pulleys up into the ceiling of the greenhouse during the day and secured to the bottom of the solar window at night. Despite considerable efforts, I was never able to unequivocally confirm a solar curtain effect. I have already mentioned the statistical difficulties associated with the measurement. In the last campaign I ran to confirm an effect, I alternated for two weeks between having the curtain down and over the solar window at night with it being up. I then examined appropriate averages. It seemed that I could detect a difference but only for the first couple of hours of the night, thereafter the results appear statistically identical. I

In private conversation with Zhang at the University of Manitoba I learned that they had tried using an internal solar curtain, but found it unacceptable. They and the Chinese solar greenhouses use an external solar curtain. The advantage of using an external solar curtain is that it will lay down on top of the solar window, while with an internal solar curtain, as I have, the curtain falls away from the window. This is important because with the ends not sealed and the gap between solar curtain and window significant convective heat losses could be significant, whereas when the solar curtain is external with the gap minimized convective losses would be minimal. The disadvantage of an external solar curtain is that it is exposed to environmental elements, in particular wind and snow. This would mean that the material used to construct the solar curtain would have to be sufficiently strong and resilient to handle these adverse conditions, and the mechanism for raising and lowering the curtain would also have to be quite stout. My plan is to make another go at it. This time constructing a horizontal thermal covering that would cover the water-filled barrels and plants, and being attached to the low south wall. Because of significant bunching, I would probably not be able to use Reflectix. I would have to use a thinner material. The aluminum in the Reflectix is particularly important in keeping the IR radiation in the greenhouse. My objective is, as always, to find a low cost solution. We'll have to see.

I have computationally verified that the water-filled barrels provide significant heat storage for night time release. I estimate that they probably add 15 degrees to the internal air temperature. The disadvantage of the 16 water-filled barrels is that they take up a lot of space. The Chinese design uses sand in the northern wall to store thermal energy. Sand is not nearly as effective as water, but by constructing an internal wall of sand there may be better use of space available. This method needs to be examined more closely and compared with the advantage of using water. In a cubit foot comparison water is superior, but there may be other conveniences in using sand or a similar material. The one disadvantage of using 55-gallon barrels is that they have a fairly low surface to volume ratio (about 0.24/inch). By increasing the surface to volume ratio both the amount of solar radiation absorbed and the amount of heat energy released into the surrounding air would increase. I plan to consider the use of long black tubing strung along the north wall. The problem with such an approach, of course, is that the volume of water would be dramatically reduced as compared to using 55-gallon barrels.

I have confirmed that covering soil at night provides a significant benefit in soil temperature. I've compared two specific methods, one using plastic laid over PVC frames, and the other using Reflectix lying low over the soil. The Reflectix is most effective, providing about a 3 degree Fahrenheit boost over bare soil. The plastic alone appears to do little, whether it is strung high over the soil or low over the soil. This is likely because the R value of the plastic is very low and it provides little or no barrier to radiative cooling. The Reflectix has a significant R value and reflects IR back toward the soil. Bear in mind that all these experiments were performed on bare soil. Were plants growing under these coverings, the results could be different. Although at night in the winter I would not expect plants to transpire much, but it would seem that covering the plants would have an affect upon transpiration and perhaps plant temperature. One might think that by just covering the plants with plastic, soil heat loss would be reduced, because it would prevent the convection of the warm air above the soil to be mixed with the colder outside air. This effect, however, appears to be small.

Probably the most valuable aspect of this research was the construction of a thermosiphon. The thermosiphon is essentially a solar water heater. It consists of a solar collector, a water reservoir, and pipes that run into the soil. A thermosiphon by virtue of the density gradients set up by temperature gradients established by the solar collector, circulates water vertically by convection through the solar collector, the water reservoir, and down through copper tubing embedded in the soil. In this way, the soil is heated. One design problem is associated with the fact that the highest temperatures in a greenhouse are located during the day at the top of the greenhouse. My concern was that the water reservoir, which is located at the highest point of the system, would become heated by the greenhouse air. Hot water located at the highest point of the system would be stable and would not convect. For this reason, I heavily insulated the water reservoir in the hope that I could insulate the water reservoir from the heated greenhouse air. The following evidences suggest that the thermosiphon is working
• the soil temperatures of the thermosiphon heated soil were often about 10 degrees F warmer than soil untreated (i.e., without any thermosiphon).
• Covering up the thermosiphon with a sheet of Reflectix yielded soil temperatures similar to that of the untreated soil.
• When the thermosiphon solar collector is covered, the thermosiphon water bucket temperature was at about the same temperature as the soil and water barrels, and much lower than the air temperatures just outside the bucket.
• Estimates of the thermosiphon water temperatures at the top and bottom of the solar collector indicate about a 15 degree F maximum difference. These estimates were obtained by taping the temperature sensors to the top and bottom thermosiphon pipes and insulating them.
• The estimated thermosiphon exit temperatures are higher than the water reservoir temperatures.
The one test that I did not perform, but am planning to, is to inject a dye by syringe through clear plastic pipe at the bottom of the thermosiphon. By measuring the time that it takes for the dye tracer to circulate through the thermosiphon, one can estimate the convection velocity.

My crude estimate is that it takes about 65 percent as much collector area as the area of soil you wanted to heat to 6 inches by 10 degrees F. I note that if you planned to keep the soil temperatures near 70 degrees F, that 10 degrees would be inadequate for South Dakota winters.

Winter Greenhouse Presentation

DISCUSSION
What seriously needs to be considered is whether a wholly passive solar greenhouse is economical. Just one gallon of fuel oil provides about 100,000 Btu. During South Dakota winters, one cannot expect to receive more than about 500 Btu/sqft/Day from solar radiation. Hence, it would take more than 200 sqft of solar collectors to make up for one gallon of fuel oil, about the size of the entire solar window. Most winter greenhouses augment solar energy collection with some other form of heating. Experiments indicate that air temperatures below 50 degrees F can be tolerated, as long as soil temperatures are kept close to 70 degrees F. Keeping soil temperatures at 70 degrees F is going to be difficult with only passive methods. This is especially the case during cloudy or overcast days. The addition of large heat storage units is, I think, essential. Water is by far the cheapest and most efficient means of storing solar energy.

One reason I began considering a system like a thermosiphon was that I realized that because the water barrel temperatures were changing very little during the night, it entailed that they were not giving up a lot of energy, and what energy they gave up went into the air. It seemed that it would be far more efficient to be able to circulate the warm water directly into the soil, thereby heating the soil up. This suggests that large thermosiphon systems might be more efficient than simply placing water barrels in the greenhouse. Most of the greenhouse loss takes place at night. So insulating plants at night should be useful. A thermal, reflective blanket that covers in a confined space the water barrels and the plants ought to be valuable if the convection losses can be reduced. The challenge is to find a means of accomplishing these ends economically.

The use of thermosiphon type systems requires a good deal of expense and effort. I have begun experimenting with something as simple as water walls (e.g. Wall O Water). The idea here is to store the solar energy close to where plants are growing simply and economically. Instead of using something like the water wall that limits solar irradiance on the mature plants because they surround the plant, I'm experimenting with walls of water in the shadow of the plants. At night the plants and water are covered with Reflectix to keep heat energy close to the plant and soil. Using this method I found that much more solar energy is being released per cubit foot of water from the water at night than from the 55-gallon barrels. This is because the surface to volume ratio is much larger. The problem with this method is that it takes up considerable soil space. One has to also worry about any shadowing of plants by the water containers. The advantage of this simple method is that instead of using an expensive and cumbersome solar curtain to cover the entire solar window and contain both plants and water barrels, the same affect is accomplished on a smaller scale perhaps more economically.

My conviction is still that a research greenhouse like the one I've constructed is necessary. Without such a facility, winter greenhouse growers have only long term estimates of the efficacy of various particular strategies. In general, these results will require examination of yield or plant growth rates. Without many seasons of experience, or operating winter greenhouses on a significantly large scale that many simultaneous experiments can be conducted, one is not likely to have sufficient parameter averaging to make justified and informed assessments. What is more, implementations of particular strategies might vary with varying implementations and circumstances. For this reason, precise and careful instrumentation is invaluable. In this regard, one might think that winter greenhouse modeling would be useful, not in producing detailed and precise estimates of heat losses, but in evaluating where monies might be best invested in winter greenhouse design. Empirical models of greenhouse heat losses have been developed (e.g., Paul Nelson, Greenhouse Operation & Management, Sixth Edition, 2003). More detailed and computational models that have been validated could likewise be useful. How much detail such models would permit is not clear, nor how much is required. In any case, such computational models have long been the aim of this project, and remain so.

PROJECT IMPACTS
I have offered to others the instruments provided by the SARE grant in an effort to evaluate their winter greenhouse projects. To date no one has taken me up on the offer. I have offered the use of my greenhouse and its instruments for local school science fair projects. To date no one has taken advantage of the offer. I have provided winter water temperature data to Chuck Waibel to aid his design of a winter solar heated fish tank. I have used my experience with winter greenhouse design to work with Jon Auer, a local builder, in the design of winter greenhouses. Finally, on January 27, 2012 I gave an hour long presentation on my winter greenhouse research at the Northern Plains Sustainable Agriculture Society annual conference in Aberdeen, SD.

OUTREACH
I have let our local high school and its science teachers know of the existence of my winter greenhouse and its measuring instruments. I have shared with our local farmer's network the existence of the same. Recently, I shared the results of over two years of data collection, modeling, and assessment with attendees at the January 27, 2012 Northern Plains Sustainable Agriculture Society conference in Aberdeen, SD. I will send along with this report a PDF version of the presentation I gave at the NPSAS conference.

PROGRAM EVALUATION
The only comment I would make regards the amount of funds requested. I was told by others who have applied for similar grants that I was insane to have requested no funds to cover the cost of my time. I requested funds to cover only the cost of instrumentation. As it turns out, I spent more money than requested on instruments, and plan to spend more even now. This is just what one would expect since it only becomes clearer what you actually need after the work is well underway. I requested no funds to cover the thousands of hours associated with the project. For one, it seems that there would be no way for SARE to be expected to cover such expenses. For another, I am retired and expect to lose money on all my many ventures. So, here is my comment. Not everyone can be expected to work under the same conditions that I do. I could find no real help to aid me in deciding how much money to apply for. One reasons that if you ask too much, you are less likely to be funded; while, one might also reason that if you request too little, the work is likely to not be done well. I have no idea what SARE's policy is in this regard. I decided to shoot for the low end in the hope of increasing the probability that I would be funded. SARE provides no guidance in this regard. Perhaps they don't wish to. One of the reasons that many people may not apply is because they think the likelihood of being funded is small. Were SARE to provide funds for a variety of funding instruments they might accommodate a more robust spectrum of clients. For example, what if there were a block of monies set aside for those who do not require “full” funding of their project? These monies would be for smaller grant amounts, but provide more funds to more people for equal dollars invested. SARE might reason that such projects are too high a risk and not likely to be efficiently spent. That is for them to decide. I suggest here two things. First, that SARE might benefit from investigating a broader spectrum of funding types; and second, that applicants would benefit from SARE's guidance with regard to funding requests.

• Northern Plains Sustainable Agriculture Society presentation by Bill Powers 1-17-12, Aberdeen, SD