Disturbance Hydro-eco-geomorphology: How do Watershed Systems Respond to Disturbances?
Sai S. Nudurupati¹ (saisiddu@uw.edu), Erkan Istanbulluoglu¹, Omer Yetemen², Domenico Caracciolo³
¹Univeristy of Washington, ²University of Newcastle, ³University of Palermo
https://gss2019-agu.ipostersessions.com/?s=32-07-8F-9E-FE-79-40-4B-EF-11-ED-DD-B7-B7-9E-B8
Presented at:

Study Area

Study area: central New Mexico. MAP=250 mm, dominated by North American Monsoon. Elevation range: 1600 to 1700 m. Vegetation: one-seed juniper trees and black grama grass coexist in NFS, and creosotebush shrubs dominate SFS. In flat terrain, shrubs dominate with sparse grass.

Figure 1: Example photo of a typical site with pronounced topography

Figure 2: Modeled spatial organization of Plant Functional Types (PFTs) from the implementation of Model A on a catchment in central New Mexico, plotted at an interval of 200 years. In this figure, we can observe that PFTs organize from an initial random condition. Trees and grass dominate north-facing slopes; shrubs cover south-facing slopes and flat areas. 

Zhou, X., Istanbulluoglu, E., & Vivoni, E. R. (2013). Modeling the ecohydrological role of aspect‐controlled radiation on tree‐grass‐shrub coexistence in a semiarid climate. Water Resources Research, 49(5), 2872-2895.

 

OPEN

Landlab - Modeling Framework

Landlab is a modeling framework developed in Python for building numerical landscape models. It provides: a 2D gridding engine to create structured and unstructured grids; data structures for storing and managing data on the grid; mappers allowing data transfer across data structures; tools for importing from⏤ and exporting to  ESRI ASCII and netCDF formats; visualization tools; and a library of components for earth surface processes. Multi-process models can be created by coupling components in a 'plug-and-play' style.

Figure 3: Structure of Landlab [Hobley et al. (2017)]

Figure 4: Schematics of Landlab Model Coupling Workflow

Figure 5: Validation of Landlab: Ecohydrology for prediction of local ecohydrologic variables.

Hobley, D. E., Adams, J. M., Nudurupati, S. S., Hutton, E. W., Gasparini, N. M., Istanbulluoglu, E., & Tucker, G. E. (2017). Creative computing with Landlab: an open-source toolkit for building, coupling, and exploring two-dimensional numerical models of Earth-surface dynamics. Earth Surface Dynamics, 5(1), 21.

Learn more about Landlab @ http://landlab.github.io/#/.

OPEN

Ecohydrologic models in Landlab

In this section, we present three models developed in Landlab to explore woody-plant encroachment.

i) Model A: CATGraSS implementation in Landlab 

In this model, local biomass production is simulated using physically based vegetation dynamics, driven by rainfall, solar radiation and temperature. Local ecohydrology model is coupled with a spatial Cellular Automata (CA) model that simulates plant establishment and mortality. Spatial CA plant competition model is executed at yearly time steps. Spatial disturbance dynamics (e.g., wildfires and grazing) is not explicitly modeled.

Figure 6: Implementation of Model A on a representative flat catchment in central New Mexico. (top) Time series of modeled vegetation cover occupied by each Plant Functional Type (PFT). (bottom) Spatial organization of PFTs plotted at an interval of 200 years. In this figure, we can observe that lack of topography favors drought-resistant shrubs. This climate is unable to sustain trees on flat surfaces. Shrubs outcompete grass.

ii) Model B: Empirical resource re-distribution based CA model built in Landlab

In this model, a stochastic CA model with two state variables, vegetation cover and soil resource storage, is used to model vegetation patterns based on probabilistic establishment-mortality interplay, mediated by post-disturbance resource redistribution. Explicit roles of climate are neglected. 

 

Figure 7: Conceptual model of states and transition processes of Ravi and D'Odorico, 2009's model.

Figure 8: Implementation of Model B for different scenarios. a) plot 1 and map 1: no fires, no grazing - baseline for sensitivity analysis, b) plot 2 and map 2: natural fire (1/10) regime and no grazing - pre-Euroamerican settlement, c) plot 3 and map 3: fires (1/100) and grazing - post Euroamerican settlements, d) plot 4 and map 4: no fires and grazing - sensitivity to fires.

iii) Model C: Model A coupled with Model B's disturbance component

This model is built by coupling Model A with Model B's spatially explicit CA disturbance component that simulates wildfires and grazing. Fire is initiated at a random cell occupied by grass and spreads to vegetated neighbors until either there are no vegetated neighbors or if the fire exceeds a specified size. Grazing converts grass to bare soil.

Figure 9: Implementation of Model C on a representative flat catchment in central New Mexico pre-European fire regime. (top) Vegetation cover occupied by each PFT is plotted with respect to time. (bottom) Resultant spatial organization of PFTs is plotted at an interval of 200 years. By comparing the ecosystem response presented in this figure with Figure 6, we can observe that the disturbance pattern used in this simulation negatively affects woody plants allowing for grass dominance.

Ravi, S., & D’Odorico, P. (2009). Post-fire resource redistribution and fertility island dynamics in shrub encroached desert grasslands: a modeling approach. Landscape Ecology, 24(3), 325-335.

 

OPEN

Ecosystem Response to Change

We wanted to explore ecosystem response to a change in disturbance pattern. Model C implementations for two such scenarios, where disturbance patterns change after 1000 years, are presented below. In these experiments, we mimic the influx of woody plant seeds into the system due to the wind by periodically seeding top and bottom two rows artificially.

Scenario 1: Natural fire regime followed by a suppressed fire regime with grazing: Seeds introduced in the boundary

Figure 10: Implementation of Model C on a representative flat catchment in central New Mexico with a disturbance pattern scenario, where the fire return period is approximately 8 years and grazing is minimal for the first 1000 years, followed by a suppressed fire regime with a fire return period of 100 years with 10% annual grazing. (top) Vegetation cover occupied by each Plant Functional Type (PFT) is plotted with respect to time. (bottom) Resultant spatial organization of PFTs is plotted at an interval of 50 years.

Scenario 2: Suppressed fire regime followed by natural fire regime

Figure 11: In this experiment, we interchange the order of disturbance regimes in scenario 1. Here, the fires are suppressed for the first 1000 years and grazing is significant. Fires intensify after 1000 years and grazing is 1%. (top) Vegetation cover occupied by each Plant Functional Type (PFT) is plotted with respect to time. (bottom) Resultant spatial organization of PFTs is plotted at an interval of 50 years.

 

Scenario 3: Natural fire regime followed by a suppressed fire regime with grazing seeds transported by wind into the domain

Now, we present another scenario where the disturbance patterns are similar to the ones considered in scenario 1, but the change in pattern occurs after 500 years. No shrub introduced on the sides, but seeds distributed by wind inside the domain.

Figure 12: Implementation of Model C on a representative flat catchment in central New Mexico with a disturbance pattern scenario, where the fire return period is approximately 8 years and grazing is minimal for the first 500 years, followed by a suppressed fire regime with a fire return period of 100 years with significant grazing, is presented. (top) Vegetation cover occupied by each Plant Functional Type (PFT) is plotted with respect to time. (bottom) Resultant spatial organization of PFTs is plotted at an interval of 50 years.

From the above experiments, we can observe that natural fire regimes with minimal grazing favor grasslands, whereas suppressed fires and intense grazing favor woody plant encroachment. We can also observe that the ecosystem state which the disturbance patterns change plays an important role in the rate of ecosystem response. 

OPEN

Fire frequency and sediment production

We use the CHILD Landscape evolution model from Yetemen et al. (2015) to investigate the role of grassland fire disturbances on the frequency and magnitude of sediment export from a semiarid central New Mexico catchment.Figure: Grassland fires consume biomass, soil production add erodible soil above bedrock, adapted from Yetemen et al. (2015). 

CHILD model is run to develop a landscape that represents similar topographic characteristics as the central NM headwater catchment.

Yetemen, O., Istanbulluoglu, E., Flores‐Cervantes, J. H., Vivoni, E. R., & Bras, R. L. (2015). Ecohydrologic role of solar radiation on landscape evolution. Water Resources Research, 51(2), 1127-1157.

Acknowledgment: This research is funded by US National Science Foundation grants: ACI-1450412 and  ACI-1148305.

OPEN

Congratulations!

Your iPoster has now been published and will be displayed on the iPoster Gallery screens at the conference.

You can keep on working with your poster. Any changes you make will be updated on the iPoster directly. You don't need to publish your poster again.

Not paid.

eLightning editing is now closed, but you can still:

• Add an Audio Presentation
• Arrange a Chat
• Change your Publishing Rules
• View Your Statistics




Please address any questions to abstracts@agu.org

00:00
Abstract

Contact Author

Print

Get iPoster

q
w
e
r
t
y
u
i
o
p
Delete
a
s
d
f
g
h
j
k
l
Enter
ABC
z
x
c
v
b
n
m
,
.
@
123
Space
#$+
Q
W
E
R
T
Y
U
I
O
P
Delete
A
S
D
F
G
H
J
K
L
Enter
abc
Z
X
C
V
B
N
M
,
.
@
123
Space
#$+
1
2
3
4
5
6
7
8
9
0
Delete
-
/
:
;
(
)
$
&
@
Enter
?
!
"
|
\
*
=
+
ABC
Space
#$+
[
]
{
}
#
%
^
*
+
=
Delete
_
\
|
~
<
>
£
¥
Enter
.
,
?
!
'
"
;
\
123
Space
ABC

Get iPoster: Please enter your email address in the field below and a link to this iPoster will be sent to you. NOTE: your email address will be shared with the author.


q
w
e
r
t
y
u
i
o
p
Delete
a
s
d
f
g
h
j
k
l
Enter
ABC
z
x
c
v
b
n
m
,
.
@
123
Space
#$+
Q
W
E
R
T
Y
U
I
O
P
Delete
A
S
D
F
G
H
J
K
L
Enter
abc
Z
X
C
V
B
N
M
,
.
@
123
Space
#$+
1
2
3
4
5
6
7
8
9
0
Delete
-
/
:
;
(
)
$
&
@
Enter
?
!
"
|
\
*
=
+
ABC
Space
#$+
[
]
{
}
#
%
^
*
+
=
Delete
_
\
|
~
<
>
£
¥
Enter
.
,
?
!
'
"
;
\
123
Space
ABC