forest/switchgrass operations

[Sheila]

I am still trying to figure out management scenarios involving
intercropping of switchgrass in pine forests. I am concerned that the
default SWAT management operations are not realistic and I wonder if
you could help me understand them. Or can some one refer me to a good
reference other than the manual.

1. From what I see switchgrass growth is initiated at the start of
the growing season (‘plant/begin growing season’). Is this correct
and do you know if it is from seed?

2. At the end of the growing season, by default, a harvest and kill
is performed. From what I understand, this means that the plant is
killed. A fraction of the material is removed and the other is
converted into residue. I don't want my plant to be killed and
removed, as it is a perennial. I wonder if I should change this to
‘harvest’ whereby plant biomass is removed (cut) without killing the
plant. Do I need to add anything else to prevent errors?

3. For trees, growth is initiated at the start of the growing season
(‘plant/begin growing season’). Is 'growth' occurring in leaves/below
ground biomass in MATURE (adult) trees? At my site I need mature
trees not seedlings.

4. At the end of the growing season, by default, a harvest and kill
is performed. Does this mean the tree goes dormant and is not chopped
down? I do not want a cut to occur.

Thanks for any help.
Sheila

[Rosie Records, CA Cooperative Fish & Wildlife Research Unit]

Hi Sheila,

I don't know about your rotations of the switchgrass and pine forest,
but maybe this can help? I've had trouble finding SWAT forestry
papers, though Gordon Putz's work might be interesting (per Jim's
recommendation). Dr. White at ARS was kind enough to advise me on
how to simulate mature pine forests in the watershed I'm working in.
I am paraphrasing somewhat, but this is basically what he told me:

1) The Harvest/Kill is a default operation assigned to all landcovers
but is NOT realistic for perennial crops/trees. Dr. White said that
the "Kill" option should be removed for all perennial crops (i.e.
trees, perennial grasses, etc.).

2) I am not sure about your questions regarding whether growth begins
from seedlings for switchgrass or mature trees for forest. However,
my assumption has been that, since the growth initiation, followed by
a harvest/kill operation, is just the SWAT default management that was
likely created originally for annual crops, that plant growth would be
initiated from seed or seedling.

3) Because I want to simulate mature pine forests also, Dr. White and
I discussed if it would be possible to set up the IGRO = 1 (i.e. that
landcover was growing at the start of the simulation). You have to
specify leaf area indices and biomass, and the literature values for
mature forest for "my" watershed were vastly larger than SWAT default
ranges for these parameters.

Dr. White told me the following:

"The inputs for biomass may not allow large enough values for this
approach to work for forest. Another way to handle this is to let the
model "warm up" for 5-10 years. Use nyskip in the .cio file to remove
these years from the output. Warm up is pretty important. I always use
at least 2 years."

Dr. White guessed that the hydrology/sediment output differences
between a 5-year and a 10-year warmup would not make much difference,
but the only way to know would be to run the models. A 10-year warmup
was prohibitively time-intensive for me to run, so I have compromised
with a 5-year warmup since the model seems to be simulating hydrology
reasonably well. My only management for pine HRUs is to initiate
plant growth at the beginning of the simulation warmup...I don't know
about using Harvest to show fall dieback but assumed that even if this
operation only simulated needle senescence, the biomass contribution
for pine would be small enough that it didn't matter anyway.

4) Regarding any timber harvest operations you might want to run, Dr
White advised:

"You could use a harvest kill to simulate forest harvest. Follow that
with a new plant operation. You may want to try this and see the
biomass change/sediment yields over time to make sure it performs as
expected."

One issue with simulating timber harvest, depending on how long of a
simulation you're running, is that it's not possible in SWAT to run a
management rotation of much more than a decade (according to Dr.
White)...so simulating timber planting, harvest, regrowth and harvest
on a 20+ year rotation may not be possible. So far I have not
simulated timber harvests in my basin because there aren't high levels
of logging and accurately assigning harvest dates and locations for
such a large watershed was too labor-intensive. I would be very
interested to hear how you decide to simulate your management.

5) Also, not sure if this helps for your intercropping, but Dr. White
also mentioned that SWAT currently does not simulate multiple
competing plants in a single HRU.

I hope this helps...Please keep me posted!

Thanks,

Rosie Records
California Cooperative Fish & Wildlife Research Unit
Humboldt State University

[dave brauer]

Shiela comments are good ones.

I suspect what one wants to model in this situation with the pine-switchgrass systems is an alley cropping practice. I am not sure that the current operation files truly have a good alley cropping scenario in them. As the pine trees mature, the annual growth of the switchgrass needs to decline to account for more solar radiation being captured by the trees. An intersting problem.. I would like to know how you end up accounting for the varibales in this situation.

--- On Tue, 1/11/11, Rosie Records, CA Cooperative Fish & Wildlife Research Unit <cac...@gmail.com> wrote:

> --
> You received this message because you are subscribed to the
> Google Groups "ArcSWAT" group.
> To post to this group, send email to arc...@googlegroups.com.
> To unsubscribe from this group, send email to arcswat+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/arcswat?hl=en.
>
>

[Sheila Christopher]

Thanks so much Rosie,

I have been long waiting for someone to help me with this issue in SWAT.  A few more questions if you don't mind based on your numbers 1-5.

1.  For switchgrass, instead of 'Harvest/Kill' then, do I use 'Harvest'  operation for switchgrass, if some of the upper biomass is taken off site after the growing season?

2.  I guess I will have to find out if SWAT grows trees/switchgrass from seed or not.  If trees start from being mature, I guess I will not have to worry about much here except to get rid of harvest and kill 

3.  What I gather here is that in the end, you chose to make your trees mature by having them grow during the 5-year warm up.  Was that enough time for them to reach maturity?  Or good enough? As you know, time to maturity could be much longer.  You did not use the method of changing IGRO to 1 and providing initial parameter values etc.  What I don't understand about your dilemma with BIO_INIT and LAI_INIT is that aren't these values on a per hectare/ratio basis, so wouldn't it not matter how large your watershed is?  My watershed is the whole southeastern US, so if this IS an issue, I guess this is not an option for me either.

You said your only management for pine was to initiate plant growth at the beginning of the warm-up simulation.  Is this simply selecting 'plant/begin .growing season'  

What happens if I want mature deciduous trees.  Then the fall die back would matter.

4/5.  My objective for now is to keep things simple and not bother with a harvest but just see what happens to the hydrology when some pine land use is replaced with switchgrass.  You are right, I can not grow two crops at once, so I am dividing the pine land use into 50% pine and 50% switchgrass

Thanks so much.

Sheila

[Jim Almendinger]

Just a note that in the manuals for SWAT 2005:
Theory, p. 330:   A Kill operation stops growth and "all plant biomass is converted to residue" -- not what you'd want for a forest, unless all growing season productivity was for deciduous leaves only (or for needle replacement in conifers). 
In-Out manual, p. 180: For a plant type defined as a "tree" (IDC = 7), new growth is partitioned between leaves/needles (20%) and woody tissue (80%).  Then a "fraction of the biomass is converted to residue" at the end of the season.  It didn't say what that fraction was...

So it remains unclear...  but at least the intent is there to include trees as a perennial plant cover.  The manuals may have lots more information -- these were just the first things I came across.
-- Jim

---

From: "Sheila Christopher" <puffchr...@gmail.com>
To: "Rosie Records, CA Cooperative Fish & Wildlife Research Unit" <cac...@gmail.com>, "dave brauer" <super_e...@yahoo.com>
Cc: "ArcSWAT" <arc...@googlegroups.com>
Sent: Tuesday, January 11, 2011 1:46:12 PM
Subject: Re: [ArcSWAT:2940] Re: forest/switchgrass operations

---

NOTE: This message was trained as non-spam. If this is wrong, please correct the training as soon as possible.
Spam
Not spam
Forget previous vote

--
Dr. James E. Almendinger
St. Croix Watershed Research Station
Science Museum of Minnesota
16910 152nd St N
Marine on St. Croix, MN  55047
tel: 651-433-5953 ext 19

[Rosie Records, CA Cooperative Fish & Wildlife Research Unit]

Hi Sheila,

sure, no problem...I will do my best to answer (sorry in advance for a
very long e-mail). I am not a hydrologist and am relatively new to
SWAT, so please take my opinions with a grain of salt. I'll try to
make it clear what is actually advice I've received from ARS and what
are my own assumptions and choices. Dr. White at ARS was really
helpful in guiding me through this. Also, he mentioned to me that
they have a plant growth expert at ARS who might be able to answer
more of these types of questions. His name is Jim Kiniry.

I embedded answers (as best I can) below:

1. For switchgrass, instead of 'Harvest/Kill' then, do I use 'Harvest'
operation for switchgrass, if some of the upper biomass is taken off
site after the growing season?

**********
I'd assume a "Harvest only" operation would make sense (mgt_op.mgt =
7). The SWAT input/output documentation says that this operation is
used for hay cuttings, where a portion of the biomass is cut and the
rest allowed to grow, so it seems that it would be applicable for the
switchgrass.

Since the documentation says that this operation harvests the portion
of the plant designated as yield, I'd guess you would want to check
what the default value is and modify if necessary. My understanding
from the '09 IO documentation is that this portion of the yield is
designated in HVSTI and WSYF--also, N and P values in the yield can be
specified (CNYLD, CPYLD). I haven't used this myself. It looks like
you can see the table of typical values in the IO documentation (it's
pages 529-530 in the version I'm using, Appendix A); I guess you would
modify the crop.dat database if these needed to be adjusted?

2. I guess I will have to find out if SWAT grows trees/switchgrass
from seed or not. If trees start from being mature, I guess I will not
have to worry about much here except to get rid of harvest and kill

***************
Please let me know what you find out about this...I had just assumed
that SWAT initiated plant growth from seed or seedlings (to me this
seemed implied in the theoretical documentation of plant growth). The
purpose of the warmup I'm running is at least partly to get the forest
landcover closer to maturity.

3. What I gather here is that in the end, you chose to make your trees
mature by having them grow during the 5-year warm up. Was that enough
time for them to reach maturity? Or good enough? As you know, time to
maturity could be much longer. You did not use the method of changing
IGRO to 1 and providing initial parameter values etc. What I don't
understand about your dilemma with BIO_INIT and LAI_INIT is that
aren't these values on a per hectare/ratio basis, so wouldn't it not
matter how large your watershed is? My watershed is the whole
southeastern US, so if this IS an issue, I guess this is not an option
for me either.

**********
These are all good questions! I asked Mike White whether a 5-year
warmup was "enough", and he basically said that you wouldn't know it
till you tried different warmup lengths and compared your outputs
under each simulation. I would feel better with a ten-year warmup (at
least), but after discovering that when I set the warmup to ten years,
SWAT was still setting up the simulation (without even running a
single year of the model) after an hour, I just went ahead with the
five year warmup. Right now I'm doing manual calibration and so such
a long simulation time isn't practical. But I should run the model
overnight at least to see how a difference in warmup affects the
outputs. I'll let you know what I find out, and hope you will do the
same.

In the process of writing this e-mail and referencing the SWAT '09 I/O
documentation, I noted that MAT_YRS (in the crop.dat database) is the
number of years required for tree species to reach full development.
For pine, it looks like this is set = 50. I wonder what would happen
if the user set this to a lower number--i.e. the number of years of
warmup? I'll ask Dr. White and let you know--if you find out first,
please let me know! In any case, it sounds like experimenting with
warmup times is the only way to really know what approach will work
best.



Regarding the BIO_INIT and LAI_INIT question, yes, I think the
BIO_INIT at least is on a kg/ha basis. What Mike White had told me is
that "The inputs for biomass may not allow large enough values for
this approach [i.e. setting IGRO=1 and allowing pre-existing landcover
at start of simulation] to work for forest." I am not sure exactly
why this wouldn't work (you could ask Dr. White) but what I found
looking through literature values for BIO_INIT is that values in the
literature that could be used for pine BIO_INIT(even at very low
levels of productivity) were orders of magnitude higher than what is
allowed for in the SWAT2009.mdb parameter ranges. So I just went with
the apparently easier method of the model warmup .

You said your only management for pine was to initiate plant growth at
the beginning of the warm-up simulation. Is this simply selecting
'plant/begin .growing season'

***************
Yep!

What happens if I want mature deciduous trees. Then the fall die back
would matter.

****************
True. I haven't used this option myself (no deciduous HRUs), but
looking at the 2009 I/O documentation, it looks like ALAI_MIN (minimum
leaf area index for plant during dormant period), BIO_LEAF (fraction
of tree biomass accumulated each year that is converted to residue
during dormancy) and BMDIEOFF (biomass dieoff fraction) in crop.dat.
It looks like DORM_HR.bsn (time threshold used to define dormancy
(hours) ) might control when your tree dormancy occurs (you might want
to read through the theoretical documentation on how plant growth is
simulated here). I would e-mail Dr. White if you want to be sure you
(and SWAT) are using these parameters correctly.

I don't think a harvest operation would be the way to go for deciduous
tree leaf biomass contribution...the default HVSTI (harvest index) for
trees is very high (0.76) and as far as I know this is total (not just
leaf) biomass.

I don't know how much more complicated you want to make your model,
but Gordon Putz and others made modifications to SWAT2005 to model
boreal forest processes, including forest litter layers and
percolation. One paper is
Watson, B. M., R. A. McKeown, G. Putz, and J. D. MacDonald.
Modification of SWAT for modelling streamflow from forested watersheds
on the Canadian Boreal Plain. Journal of Environemtanl Engineering
Sciences 7:145-159. I don't have a digital copy but could send you a
hard copy if you'd like.

Also, someone posted about deciduous forest modeling in KY--maybe
you've already seen this or been in touch with CJ? See the thread at
http://groups.google.com/group/arcswat/browse_thread/thread/2e9b3961491c2bdf/dea7c2bf08eabace?lnk=gst&q=jim+forest+deciduous#dea7c2bf08eabace
.

Good luck! Keep me posted, and I'll let you know what I found out
from varying warmup times, and any more information on time to
maturity for forest.

[Sheila Christopher]

Hi Jim,

Thanks.  I saw those too.  Definitely would not want harvest and kill if everything (including bowl and roots) goes to residue.  I do need to figure this out.  

Also, given the second point  you made about how trees are handled, it makes me suspect that tress ARE mature at the start of simulation.  Also the fact that trees are allowed to go into dormancy such that " leaf biomass is converted to residue and the LAI for the tree species is set to the minimum value allowed" I need to ask some people in Texas!

I will keep everyone posted

S

[dave brauer]

Having worked in agroforestry for over 10 years, I do not think a simulation in which the pine land use is divided randomly 50% to switchgrass and 50% to pine will simulate an alleycropping/inter cropping practice, unless the size of the HRU is small.   Based for the above statement is there are  numerous reports that the productivity of either component is affected by the arrangement (spatially) of  the two components. Such refrences can be readily found in the 2nd edition of North American Agroforestry (ASA publication).   The closest land use/ cropping system may be a brushy rangeland. It may be possible to modify the operations of this management practice to better approximate a pine swichgras intercropping. One problem that is going to be encountered is that over a long simulation the productivity of the switchgrass and pine are going to change.   An interesting problem --- On Tue, 1/11/11, Sheila Christopher <puffchr...@gmail.com> wrote:

From: Sheila Christopher <puffchr...@gmail.com> Subject: Re: [ArcSWAT:2940] Re: forest/switchgrass operations To: "Rosie Records, CA Cooperative Fish & Wildlife Research Unit" <cac...@gmail.com>, "dave brauer" <super_e...@yahoo.com> Cc: "ArcSWAT" <arc...@googlegroups.com>

[Sheila Christopher]

I hear you Dave, but there is no way to grow 2 crops in an HRU in SWAT, even if it is rangeland.  THe APEX model may offer some way of growing 2 crops and them importing it into SWAT, but I have not explored that option.  THe other option is to create a new land use called intercropping and insert the parameter values that we are getting in the field.

[Sheila Christopher]

Just found out my trees need to be 2-4 years old, not mature, but this would still involve some sort of adjustment of the model.  I will keep you posted.

Sheila

[Johanson, Christopher]

I've been doing some experiments.

For FRSD, changing MAT_YRS from 50, to 10, to 5, to 80 makes no difference in Biomass, LAI, or YLD (or any outputs that I can see). Not sure why I'm getting a YLD, as the only operation I have is to Plant in year one. No harvest or Kill. I can only assume the YLD is perhaps a theoretical yield? MAT_YRS does not seem to be functional for me.

Also, my FRSD does appear to grow from seeds, however it seems to grow unrealistically fast for the first 2 years, going from 0 biomass (metric tons per hectare) to almost 25 metric tons per hectare in month 24.

After 18 years my FRSD has a biomass of 40 metric tons/hectare. I'm going to try a longer simulation on a smaller watershed to see if my FRSD will continue to grow to what I'm told is reasonable mature biomass of approx. 300 metric tons/hectare.

CJ Johanson
________________________________________
From: arc...@googlegroups.com [arc...@googlegroups.com] On Behalf Of Rosie Records, CA Cooperative Fish & Wildlife Research Unit [cac...@gmail.com]
Sent: Tuesday, January 11, 2011 3:59 PM
To: ArcSWAT
Subject: [ArcSWAT:2942] Re: forest/switchgrass operations

Hi Sheila,

Thanks,

--

[Bini]

Hi Sheila,

I have the same question as you have about the growth of trees in
SWAT. Do you manage to get answer for your question? If you find any,
please post it.

Bini,

On Jan 12, 9:08 pm, Sheila Christopher <puffchristop...@gmail.com>
wrote:

> >http://groups.google.com/group/arcswat/browse_thread/thread/2e9b39614...

> > .
>
> > Good luck!  Keep me posted, and I'll let you know what I found out
> > from varying warmup times, and any more information on time to
> > maturity for forest.
>
> > Thanks,
>
> > Rosie Records
> > California Cooperative Fish & Wildlife Research Unit
> > Humboldt State University
>
> > --
> > You received this message because you are subscribed to the Google Groups
> > "ArcSWAT" group.
> > To post to this group, send email to arc...@googlegroups.com.
> > To unsubscribe from this group, send email to

> > arcswat+u...@googlegroups.com<arcswat%2Bunsu...@googlegroups.com­>
> > .

> > For more options, visit this group at

> >http://groups.google.com/group/arcswat?hl=en.- Hide quoted text -
>
> - Show quoted text -

[Johanson, Christopher]

Hello,

Here are some results from my experiments on one FRSD hru.

Monitored effects of changes to various input parameters on one FRSD HRU on BIOMt_ha, LAI, and YLDt_ha.

MAT_YRS has no effect on monthly output.
Monthly output is identical with harvest/kill operation or harvest only operation.
BMX_TREES has no effect on output. When set to 50, model will continue to grow right past 50, at the same rate as when it is set to 100, 200 or 300. When set to 300, even after 1000 years BIOMt_ha is only 128.
A Harvest Only operation scheduled for year 20 does reduce the biomass, but not as much as expected.
CHTMX has very little effect on monthly output. In fact, when changed from 20 to 6, there was slightly more BIOMt_ha after 50 years.
Creating an initial crop of FRSD has a small initial effect, but not as much as expected, and the default model grown from seedlings catches up in about 25 years.
YLDt_ha is present in monthly output even when there is no harvest operation specified. YLDt_ha is 0 in yearly output when there is no harvest.
BIOMDIEOFF has no effect on monthly output.
BIO_LEAF does have an effect. The default value of 0.3 allocates nearly a third of biomass as leaves falling from the tree each fall.
Simulating a transplant by setting CURYR_MAT from 0 to 20; LAI_INIT from 0 to 5; BIO_INIT from 0 to 200; CNOP from 0 to 40 creates no change in yearly output.
On a 50 year run with yearly output, when we pull out the sediment yield from each FRSD hru and average it, the value seems slightly high, but pretty close (according to Summary of Sediment Yield Data From Forested Land in the United States; Patric, James, H. Journal of Forestry, February 1984, pgs 101-104).
Changing the SCS curve numbers in the crop database does not affect the model. Perhaps because these numbers are 'set' during hru creation?

My conclusion is that if you are trying to simulate tree growth and forestry operations, it's not too realistic. But if you just look at the hydrology, and allow some warm-up time, the model works pretty well.

CJ Johanson

________________________________________
From: arc...@googlegroups.com [arc...@googlegroups.com] On Behalf Of Bini [bini...@gmail.com]
Sent: Wednesday, February 09, 2011 11:24 AM
To: ArcSWAT
Subject: [ArcSWAT:3061] Re: forest/switchgrass operations

Hi Sheila,

Bini,

To unsubscribe from this group, send email to arcswat+u...@googlegroups.com.

[Sheila Christopher]

Hi all,

Thanks for this awesome discussion on how trees and swtichgrass are handled in SWAT.  I have been away from the User Groups for a few weeks now and am getting back into it!

After discussion my project with my colleagues, we decided, that you really can't not model interceropping in SWAT, as Dave indicated, so I am not focusing on that right now...until we can create a land use coverage called 'intercropping!'

What I am focusing on is getting my model running with a realistic pine land use.  I am STILL struggling to find the answer to how pine is treated in SWAT,  I did correspond with a friend who works closely with Srini and he did offer this info but I still am a little skeptical about #1:

1.  Pine is mature at the start of the simulation  (all trees are the same age)
2.  By default, pine growth is initiated at the start of the growing season (‘plant/begin growing season’)
3.  At the end of the growing season, by default a ‘harvest and kill’ is performed.  What this means is the trees go dormant and are not actually removed from the watershed.  Leaf biomass is converted to residue. 

What I would like to have in SWAT is 3 size classes, something like 0-4, 4-10, and 10-30 and I need to figure out how to do this.  I don't want all of my trees the same age at the start of the simulation!  This is not realistic.  I am not sure how to do this.  Any thoughts?  Would this involve creating 3 separate Pine land uses?  Would it then involve Editing Management parameters.  And then, in this window, specifying Initial Land Cover such that the initial land cover would be different for each of the three pine land use covers (resulting in 3 different size classes).  

Happy Swating.

Sheila