Modeling spatial variability of SOM in forested ecosystems cannot predict changes or storages in SC

1
Modeling spatial variability of SOM in forested
ecosystemscannot predict changes (or storages)
in SC at small scalesdont forget landscape
position

• St Paul meeting
• January 17-18, 2006

2

• Scale
• Stand to mountain scale
• Within similar climatic and forest conditions
• Varying edaphic and topographic conditions
• Why
• Management at stand level
• Most SOM models have spatial r2 lt 0.2
• Ultimate Goal (SOM site indices)
• Develop methods to predict site specific SOM
  storage and accumulation
• Given site conditions what is steady state SOM
  pool
• Given SOM at To what will be SOM at Tn

3
Geomorphic-Soil Catena approach SC models f(
climate driven input decomposition) Decom.
f( edaphic, grainsize..landscape position)
Accum. f( lateral transport. landscape position)
Soil moisture Grain size Texture SOM,
cations Slope Slope length Flow Ac
4
Method Given these site conditions what is SS
SOM ?

• Data from long-term studies
• Green Mts Vt, Post and Curtis, Johnson
• Bisley, Puerto Rico, Scatena, Johnson
• Adirondacks, NY Heimburger, Johnson
• Tier 3 plots ???
• Model using site conditions
• GIS based regression slope, FAC
• Century with Spatially varying parameters
• Extrapolate using Spatial Statistics and GIS
• Site indices

5
Landscape approach

• Landscape attributes (constant climate,
  vegetation)
• Gessler et al 2000 California hillslope
• 70-80 F(slope and catchment area)
• Chaplot et al 2001 French agricultural fields
• 75 f(Elev. above channel, slope, contributing
  area)
• Van Oost etal 2005 Erosion and SOM, Danish Ag.
• Thompson Kolka 2005..
• Soil Series correlations
• Present is key to future
• Knops and Tillman 2000
• Examples

6
Simulated vs observed SOMCentury Based Luquillo
Mountains
1-1
Across Forest Types r2 0.7Within Forest Types
r2 0.2 -.35
7
Century Based Luquillo SOMVegetation type
elevation
8
SOM vs Distance to Stream
SOM vs Distance to Ridge
5 m
10 m
R2 0.35
R2 -0.37
9
Green Mountains
Post and Curtis 1970 24 sites across Green
Mts 70-80 years old sites Century
Pnet Century GIS
10
SOM 0-20 cm(excluding forest floor)

• Assumptions
• Variable climate
• Uniform topography
• Uniform vegetation
• Improvements
• Deeper is better
• Site specific inputs

11
SOM FF
Where are these High storage sites?
12
Stepwise regression without GIS or Century
Elevation Depth Latitude Moisture Slope
13
Conclusions

• Landscape regressions gt Century
• Include landscape position
• Regional or Forest type based relations
• Depth, grid size for landscape features
• Estimate maximum storage vs accumulation

SOM Storage
F to F
Sm
NF to F
T
14

• Family of curves based on site conditions
• How far is my site from Smax

F to F
Sm
NF to F
SOM Storage
T