I have used Simforest to create several simulations on two sets of date comming from two different terraces in Zoar Valley. I used all of the trees in the sampled quadrats to build a forest and then subjected it to all the different site treatments. Afterwards I exported them to Excel, determined what year each tree was born using a subtraction equation and then seperated out the trees that were still alive when the simulation stopped. My plan is to now run statistical tests on them data using SPSS.

~Erin K. Pfeil

SimForest Issues

I was attempting to run the SimForest simulation to find the number of tree species using the Zoar Valley Species Table, but I also encountered the same problems as Erin did, namely that this new species table is somehow incompatible with the equation list. Since the program does not run correctly when you modify the species table and the equation list at the same time, it will hamper our ability to track the number of trees of different species growing in Zoar Valley. 

I am also working to add the available light equation to my NetLogo program. This will use the leaf weight equation, which we have defined for the time being as being equal to 0.4 (the constant) * diameter of the tree.  

- Irena

I still cannot get Simforest to run all of our customizations at once. It will only cooperate with one or two at a time depending on what you want it to do. I emailed the creators of simforest and I was told that the project was no longer active and that they had no advice. This is very frustrating for me, I am not able to continue with my part of the project. I tried to add the number trees equation parameter to the original species table and then build the Zoar Valley species composition but it still won't work. I am going to have Irena try to run the new parameter with the zoar valley species table and see if it works for her. I am curious to see if it just my program or java not working correctly or if its an inherent problem with the program all together.

~Erin K. Pfeil

NetLogo Update

I've had some progress on the NetLogo model in that I finally have the trees growing to reflect their changing height and diameter. I am currently working on getting the user to insert their own tree species via a drop-down menu. This has been set up in the program, but a problem has been encountered in that the trees are currently only growing with the current values that the program was designed around. Now, the program must be able to work with a wide variety of species, so this is the last major issue that needs to be addressed. I think that this could be due to the fact that the available light equation is not reflecting the changing tree types, however, Erin is trying to address that by emailing a few of the people who worked on the SimForest project, and while some are no longer actively associated with it hopefully she can come across someone who will provide some info regarding the Botkin equations being used. 

After this (fairly major) issue has been addressed, all that remains will be to attempt to run the program with an eye to how it reflects the actual composition of Zoar Valley. 

Ok, I think I found a way to make it work. If you add the parameter and the equation to the original species table then change the species composition of the table it seems to work. I am going to print out the table that Dr. Lazar recently created and after adding the parameter "num trees" I will build the Zoar Valley species table. Cross your fingers everybody! I will post whether or not I am successfull later tonight.

~Erin K. Pfeil

Well Im still trying to get Simforest to work and produce usable data. After many hours over several days, I got it to accept the numbertrees equation into the Zoar Valley species composition table, but now the parameter won't actually do anything. Hey what a surprise. I will try to start all over again and see if it will work. It seems Simforest has serious issues allowing you to use more then one custom table at a time. If I use the new equation with the original tree list, it works fine and vise versa. When I try to use all of our custom stuff together it fails, several times its actually crashed the program. If anyone has any suggestions, let me know.

~Erin K. Pfeil

OK, I found some things that might be useful:

Ci = A constant of proportionallity relating leaf weight to tree diameter.

SLA = actually the leaf weight above tree n, the sum of the weight of the leaves on all trees taller then tree n.

soooo... when i get my simforest working again (which I optimistically am hoping is before our meeting today) i will see what equations they give in their equation list. I'm assuming that SLA depends on values provided from other equations, particularly one or ones that deal with leaf weight.

~Erin K. Pfeil

I have recieved the Forest Dynamics book from Ohiolink and found the equation for light availibility. Irena has put it into the following format with values I found in the book:

AL=1e^1(-(1/6000)*SLA)

The only issue at this moment is what value to give to SLA (Shading leaf weight). Three assumptions are given for the this variable:

1.) all of the leaf wieght is concentrated at a point at the top of a tree.
2.) therefore that all of the leaves of one tree shade any shroter tree and...
3.)... conversely, that none of the leaves of any shorter tree shade any leaf of a taller tree.

It states that SLA is leaf weight, which would differ from species to species, particularly as stated by Botkin, shaded tolerant and shade intolerant species. Earlier in the book its states that leaf wieght is the square of the diameter of the leaf, and an equations is given:

W=CD^2 (C is a constant with an i as subtext, which and what its value is, I dont know)

I am still working on getting my Simforest to work. My computer decided to bail on my over the weekend and I had to fix it and reinstall vista. Now I am in the process of finding a working download of Java 1.2 so that I can at least get Simforest up and running again, since it appears I no longer have that version of Java and the program won't open at all. Perhaps it will run correctly after I download Java again.

~Erin K. Pfeil

NetLogo Update

So far, the NetLogo model that we currently have plants one tree species. It has both birth, recruitment and death of the tree. The user can plant the tree anywhere using the cursor. There was a slight problem with recursion being too extreme in the program, as there were three equations (Growth, Maximum Potential Growth, and Diameter) which depended on each other and as a result couldn’t properly function when running at the same time. This was corrected. Now, the next goals we have for the model are:

                - Changing the size of the trees in the model to reflect their growth

                - Planting trees of different species

                - Adding graphs that reflect the changing state of certain tree variables

 

-- Irena

 

I have constructed a site table to reflect the affect of global warming on eastern hardwood forest, particularly Zoar Valley. I used data I found on this website:

http://www.epa.gov/climatechange/science/futurecc.html

I will make another that reflects a more extreme increase in temperature. This table shows the minimum predicted difference (1.8 degrees Celsius). There are no reliable predicted changes for rainfall, so I doubled it for every month. My guess is as good as any at the moment. I have also created a working excel notebook that contains tree table data from the historic site table and the global warming site table. I do not know what to really do with it at this time, until it can be broken up into more manageable pieces.

Once my Glass Box program is repaired (again) I will build the exact terraces I have data for as they exist. I feel this will offer an even more realistic simulation.

I have also ordered a book (Forest Dynamics: An ecological Model, by Botkin) through Ohio Link that Irena found, it should be very useful.


~Erin K. Pfeil

NetLogo Model Progress

I have been working on some of the issues with the NetLogo model and made the following changes:

 - the waterstress variable has been changed to 0.1 to produce a high enough value in the sitequality equation to ensure that trees are being recruited. 

- availablelight has been changed to 0.99 (from 0.74), which now allows recruitment to proceed as well

With these variables, trees are now both dying and being recruited. 

Next on the list is working on height, diameter and age. I am working towards having the entire model based on "ticks", which is simply a different method of representing the passage of time - the model changes after a certain discrete amount of time (one "tick") rather than continuously. This could help reflect the way that a change in a tree is shown after every whole year, the way it is in other models. 

Irena

I have create two site table files, one for 1890 and one for 1986. I will use both files and data given to me by Dr. Diggins to try to construct quadrats on simforest. I will continue to search for more up to date weather data and contiue my literature searches.

~Erin K. Pfeil

I'm not entirely sure what to do with the equations. Water stress is low for Zoar Valley, so you could try changing the value so that it is smaller then .4 or if that makes it worse experiment and try a larger number until the equation works. Since its very difficult to inquire about these things to the creators of the simforest program, that is what i would do. Perhaps if you can get it to work using a different value, we can figure out the meaning of the value itself or perhaps there is a problem somewhere else in the equation. I also suggest raising the amount of available light, if I understand the previous blog entry correctly. We may be underestimating it far more then we thought. If the equipment was available, it would be best to find the available light ourselves, but maybe I can find a value from a study done on a similar old growth forest.

Regarding the excel file, I have looked at it and it does not have data for 2006. Soooo i will use the newest one I have, to create the new site quality table. I am also in the process of emailing the site quality table I have already created and send that along as well.


~Erin K. Pfeil

Progress on NetLogo so far

After going through the equations used in NetLogo and comparing them to those used in SimForest, I saw that for the most part these equations are functioning and being used correctly in combination with other equations. Currently, I am going through the values generated by these equations to see if the values being generated are actually correct.

I’ve found a few issues with the equations after going through them, mostly related to the recruitment equation, one of the three major equations used to drive the model (the other two being birth and death). Recruitment determines the number of new saplings planted depending on a particular species’ tolerance to light/shade (represented in the program on a scale of 1 – shade-intolerant, 2 – shade intermediate, or 3 – shade-tolerant). Two parts of the main recruitment equation involve the variable sitequality, which is itself an equation involving three other variables: wiltfactor, nitrogenfactor and tempfactor.

One problem that arose involves the wiltfactor equation, which incorporates the variable “waterstress”. This variable is defined in the SimForest equation editor as “[returning] the water stress at the site in the current year, computed based on the site's properties using a complex equation”. However, no other information is given on how to actually calculate the value. This creates a problem as the value we are currently using for waterstress in the wiltfactor equation (arbitrarily chosen to be 0.4) results in the equation producing a value of zero. This occurs because the wiltfactor equation is designed to return the higher of two values, either 0 or a value produced by list ((1.0 - ((waterstress / wlMax) ^ 2))) (0) In this case, the value that comes from this equation is negative, so 0 is always chosen. This in turn leads to the entire sitequality equation to be zero, as sitequality is defined as

tempfactor * nitrogenfactor nitrogen * wiltfactor

Some value needs to be found for waterstress that allows wiltfactor to return a non-zero value, so that it can in turn be used in sitequality

Another second problem that arises is with the variable representing available light. This calculation represents the amount of light available at the current tree's height, and the creators of SimForest admit that they have not included the actual equations for inspection in either their Glass Box or Black Box versions, which makes it difficult to determine if the value chosen is correct. In this case, we have again chosen an arbitrary value, 0.74, for available light.

This is a problem when we get to the recruitment equation. When light = 1, the formula availLight >= 0.989999 and sitequality > 0.0 is used to determine whether or not to perform the recruitment equation or just set recruitment to zero. Since 0.74 is always <>

A similar problem arises when light = 2. The equation used is minSaplingLight > availLight and availLight < light =" 2.

The two values for available light needed for light = 1 and light = 2 are not compatible, but the available light must be calculated so that one of these equations does not keep constantly returning 0 for tree recruitment.

NOTE: light = 3 has not been discussed yet because it relies upon the sitequality equation, which is currently returning 0.

Resolving these two issues is important in allowing development of the model to continue. In the meantime, I am working on adding the rest of the equations that the model uses (Growth, Age, etc), particularly those pertaining to the growth of the tree itself, which I am planning on incorporating into the NetLogo model if possible.

Forest Dynamics

http://research.microsoft.com/displayArticle.aspx?0rc=n&id=2029

Weather Data

Hello Everyone

Here is a site were I finally found weather data that begins in 1900 and runs up to 2006.

http://cdiac.ornl.gov/epubs/ndp/ushcn/monthly.html

I chose the city of Brockport, New York. It is ~96 mi from Gowanda, but it is the closest and what I feel more similar to Zoar Valley then the given New Hampshire site parameters. I created an excel spreadsheet (through the use of the site) and highlighted which columns we will probably be using. The only thing I am having problems with is converting the data from inches to millimeters in excel and also Fahrenheit to Celsius. Can anyone tell me how to do it? I will email you the file.

~Erin K. Pfeil

Goals for NetLogo Program

Now that we are aware of the different capabilities of both the Glass Box and Black Box versions of SimForest, we are also looking to re-create some of the capabilities of these programs in NetLogo. This software will allow us to create a forest-modeling program from the ground-up, and perhaps give a little more flexibility in what we can do with either of the previously-mentioned versions of SimForest. However, doing this presents some unique challenges, since we are starting from scratch. Using the SimForest programs as a basis, we must translate their equations, which are done in Director Lingo, into the language of NetLogo, which has its own unique syntax, procedures and commands. Some of the goals to work towards are:

· getting the recruitment equation running

· incorporating different species into the simulation

· being able to plant the trees one by one

· introducing a topographical map of Zoar Valley into the simulation background

Once a running simulation has been achieved, it can be cross-checked against the other models to check if it is producing realistic values.

Simforest Equation Editor

We are going to look at the SimForest glassbox simulation Equation Editor:
http://ddc.hampshire.edu/simforest/software/docs/eq-editor-tutorial.html

Netlogo Simulation

Irena,
Please upload the latest version of the Netlogo simulation on the YSU MRO-W webpage.

Manipulating Equations

These are the equations that I want to try to manipulate, in order to simulate beech bark disease:

AGE = (age(treetable)+1.0)
DIAMETER = (diameter(treetable+growth)
GROWTH = (limiting factors x maximum potential growth)
LIMITING FACOTRS = (light factor x site quality)


Beech bark disease appears to affect trees that have high nitrogen levels more then trees that do not. Beech trees of Zoar Valley would most likely have high nitrogen levels, based on a paper I read that stated that trees in old growth areas (these systems are more productive) have higher nitrogen. However, I do not think that this is something that can be manipulated in simforest. Beech is set to be nitrogen tolerant (2.0), I think this will have to do for now.

Another problem with changing the equations is that the site quality table is somewhat of a mystery. I have no idea what units any of it is in, or where any of the averages came from. I will look at the simforest website to try and figure this out.

It will probably be easiest to limit the age of beech rather then the size. In the program size (diameter) depends on the site quality and light, which would change the actual growth of all of the trees in the simulation, which is not what we are trying to do. Site quality may be something we want to change for a different affect thought, and I would really like to simulate actual Zoar Valley conditions and not those of New Hampshire, since I'm sure they differ.

~Erin K. Pfeil

Here are some ideas for altering parameters to mimic beach bark disease and emerald ash borer:

~ I have recently read in a long term study on hardwood forest regeneration that when mature american beech trees are about to succumb to beech bark disease they will send up alot of root shoots. To account for this, the life span of beech will have to be shortened (by half or so), but they will also somehow have to be made more prolific. It would also be good to know if beech "behaves" this way anytime it is about to die, not only when it has beech bark disease.

~ Emerald ash borer can be simulated by changing the death equation, or by greatly increasing the chances of an ash tree dieing. The literature states that emerald ash borer can cause mortal damage to an ash tree within 3 - 4 years of the tree encountering the insect.

I will post more possibilities later,

~ Erin Pfeil

Within the past two weeks I created a poster that was presented at YSU's Biology Days. It turned out a well and won one of the second tier prizes for the event. It will probably also be presented this week for an event showcasing the STEM college at YSU. As far as the research, I have been compiling journal articles so that we may accurately and realistically alter parameters to reflect conditions such as emerald ash borer and hemlock wooly adelgid. Right now I am focusing on those two pathogens. Tommorow I will post my thoughts and ideas on how we might use glass box and netlogo to accomplish this.

~Erin K. Pfeil

Over the past several weeks a lot has been accomplished on our forest modeling project. We are still working on getting the Black Box version of SimForest to function, but Glass box is working perfectly. I have constructed a species table that is representative of Zoar Valley and some of its local conditions (such as the water table depth). Parameters such as average rainfall per month and temperature are still being decided upon. To test future conditions and affects on forest composition, rainfall and temperature will be adjusted several times, so this is taking additional research and consideration. My literature searches are ongoing to find information about our parameters and environmental variables we need to change and adjust to raise the predictive power of our models.
We have presented our research so far at QUEST, which is a conference put on by YSU for student research. I was responsible for the research site introduction and for an explanation of how we were going to use existing and published data sources to model forest composition. Next week on the 17th, I will be presenting a poster of our research at YSU’s Biology Days departmental conference.

~ Erin K. Pfeil