Experimental design
Our experimental design consisted of 15 - 1 ha blocks spatially
distributed across the study area selected to have similar total
percentage vegetation cover within a block, but to vary in the
proportion of perennial grass and mesquite cover among blocks. At
each block location, four 15m x 25m plots were identified with
similar percentage grass and shrub cover with a focal patch in the
center. Plots were arrayed perpendicular lengthwise to the
dominant southwest to northeast wind direction, and were >15 m
apart to minimize between-plot treatment effects.
Each plot within a block was randomly assigned to one of four
connectivity treatments: (1) plant scale where all mesquite plants
within and surrounding the focal patch were killed in place to
modify competitive interactions between woody plants and recovery
of perennial grasses and other herbaceous plants with no direct
effects on horizontal transport by wind and water, (2) patch scale
where Connectivity Modifiers (ConMods, Okin et al. 2015) were
located in bare soil interspaces between plants in the focal patch
to reduce gap size and to modify transport of water, soil,
nutrients, litter, and herbaceous seeds, (3) both patch- and
plant-scale manipulations were conducted in each focal patch, and
(4) no manipulations [controls]. Block and plot selection were
completed in June 2012 followed by the characterization of initial
vegetation cover in all plots in June 2013 when treatments were
initiated.
Plant-scale manipulations (+WPI)
At the plant scale, repeat photos of ten randomly selected ConMod
microplots and ten randomly selected control microplots were taken
annually at peak growing season (September) from 2013 to 2021.
Standardized overhead photos were used to estimate perennial grass
cover as our response variable, and cover of forbs (annual and
perennial), annual grasses, and litter as fine-scale factors that
could influence grass cover by providing favorable microhabitats
for grass seedlings (litter) or by competing with perennial grass
plants (forbs, annual grasses). Overhead photos were collected
from 2013 to 2021, and image analysis software was used to
classify 100 random points by species, soil or litter (USDA
SamplePoint software; Booth et al. 2006, Peters et al. 2020).
Species cover in each year was summed to obtain functional group
cover: annual grasses or forbs.
Patch-scale manipulations (+WPa)
Lateral photos obtained from the same microplot locations as
overhead photos were used to estimate the redistribution of soil
and litter by wind and water from bare interspaces to ConMods or
to herbaceous plant canopies. Lateral imagery was collected from
2013 to 2017 at ground level from each cardinal direction to
estimate the vertical accumulation of soil and litter (hereafter
"litter"). The area between the tops and sides of each
ConMod’s outer rods and the top contour of the combined litter and
soil surface was determined (methods described in Peters et al
2020; SigmaScan pro 5.0: Systat Software, Inc. San Jose, CA USA)
using Trace Measurement Mode with Area and Distance measurement
options. The area difference between years for the same ConMod was
a measure of the change in vertical accumulation of litter or
soil. Lateral photos were discontinued after 2017 when density and
cover of herbaceous plants made the analysis of litter and soil
accumulation difficult and prone to error.
Landscape-scale factors
Factors at the landscape scale were initial shrub cover based on
classified imagery from the Quickbird satellite mission in 2011
and water year precipitation. Quickbird source imagery used in the
classification should be available at
https://earth.esa.int/eogateway/catalog/quickbird-2-esa-archive.
Cover classification into bare ground, herbaceous, and shrub
categories utilized a supervised classification tool in ArcGIS’s
Spatial Analysis extension. Training data were provided by field
staff who identified and georeferenced representative samples of
each cover type. ArcGIS was then used to analyze spectral
signatures of the QuickBird imagery and classify pixels.
Daily precipitation data from 13 weather stations spatially
distributed throughout the study area were summed for the water
year (1 October to 31 September). For each block, precipitation
from the nearest weather station was used in the analysis. These
weather station datasets are referenced in the Data Provenance
metadata element.
References
Okin, Gregory S., Mariano Moreno-de las Heras, Patricia M. Saco,
Heather L. Throop, Enrique R. Vivoni, Anthony J. Parsons, John
Wainwright, and Debra P. C. Peters. 2015. "Connectivity in
Dryland Landscapes: Shifting Concepts of Spatial
Interactions." Frontiers in Ecology and the Environment 13
(1): 20–27. https://doi.org/10.1890/140163.
Booth, D. Terrance, Samuel E. Cox, and Robert D. Berryman.
"Point sampling digital imagery with 'SamplePoint'."
Environmental Monitoring and Assessment 123 (2006): 97-108.
Peters, Debra P. C., Gregory S. Okin, Jeffrey E. Herrick, Heather
M. Savoy, John P. Anderson, Stacey L. P. Scroggs, and Junzhe
Zhang. 2020. "Modifying Connectivity to Promote State Change
Reversal: The Importance of Geomorphic Context and Plant–Soil
Feedbacks." Ecology 101 (9): e03069.
https://doi.org/10.1002/ecy.3069.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.
This method step describes provenance-based metadata as specified in the LTER EML Best Practices.
This provenance metadata does not contain entity specific information.