low_impact_lot_practices Materials supporting Voter, C.B. and S.P. Loheide II.
Urban Residential Surface and Subsurface Hydrology: Synergistic Effects of Low-Impact
Features at the Parcel-Scale. Water Resources Research, 54(10):8216-8233.
https://doi.org/10.1029/2018WR022534). This is a copy of the git repository:
https://github.com/cvoter/low_impact_lot_practices#low_impact_lot_practices
data
data/colormaps
Includes several matlab colormaps used in plots. All but map_ylgrbu.mat and
map_reverse_jet.mat are tied to values of parcelCover (i.e., for values spanning 1
through 9).
data/initial_pressure
Includes initial pressure (m) conditions for silt loam (SiL),
moderately-compacted silt loam (SiL2c), and highly-compacted silt loam (SiL10c)
baseline soil conditions. Each *.mat file includes the following variables:
pSP: complete record of pressure head (m) on April 1 for 300 year spinup
simulation
pSP30: record of pressure head (m) on April 1 at end of each 30-year loop of
weather (i.e., every 30 years)
sSP: as pSP, but saturation (-)
sSP30: as pSP30, but saturation (-)
spIC: pressure head on April 1 of last simulated year, used as initial
conditions for model
data/layouts
Includes information about 2D layout of parcel features and microtopography
elevations. Files include:
lot_microelev.mat: deviations in elevation for microtopography scenario.
Variables include: microElev: matrix (ny X nx) indicating deviations in elevation
(m) from overall land slope (2%) for microtopography scenarios. sd: standard
deviation in elevation specified during random generation of elevations (Onstad,
1984) RRcalc: calculated random roughness from randomly generated deviations in
elevation (Onstad, 1984) DScalc: calculated depression storage (cm), based on
calculated random roughness (Onstad, 1984)
Lot00.mat: limited domain information for lot with connected downspout and
connected sidewalk. Variables include: dx,dy: discretization in x and y direction
(0.5m for both) nx,ny: number of elements in x and y direction (44 and 75,
respectively) xL,yL: lower limit of x and y domain (0m for both) P,Q: number of
processors allocated to x and y domain (4 and 5, respectively) parcelCover: matrix
(ny X nx) indicating type of feature on each pixel. Key: 0 = turfgrass, 1 = street,
2 = alley, 3 = parking lot, 4 = sidewalk, 5 = driveway, 6 = frontwalk, 7 = house, 8
= house2 (extra house behind garage), 9 = garage fc: coordinates of each impervious
feature. Rows correspond to parcelCover key (i.e., row #1 = street coordinates).
Columns indicate [lowerX upperX lowerY upperY].
Lot10.mat: same as Lot00, but with disconnected downspout
Lot01.mat: same as Lot00, but with disconnected sidewalk
Lot11.mat: same as Lot00, but with both disconnected downspout and disconnected
sidewalk
data/model_inputs
Subdirectories indicate model runs, formatted 'Lot%d%d%d%d_%s_%s', where the
four integers indicate presence (1) or absence (0) of disconnected downspout,
disconnected sidewalk, transverse slope to front walk and driveway, and
microtopography; the first string indicates soil type (SiL, silt loam; SiL2c,
moderately-compacted; SiL10c, highly-compacted); and the last string indicates the
growing season weather scenario (average or dry). Each directory includes the file
domainInfo.mat, which includes the following variables:
dx,dy,dz: discretization in x, y, z (0.5m horizontal, 0.1m vertical) nx,ny,nz:
number of elements in x, y, z (44, 75, 100) x,y,z: vectors with the center x, y, or
z coordinate of each element P,Q,R: number of processors in x, y, z (4, 5, 1)
domainArea: surface area (m2) of domain elev,DScalc: final elevation (m) for each
pixel and calculated depression storage based on random roughness approach (Onstad,
1984) Ks_imperv,porosity_imperv,Sres_imperv,Ssat_imperv,VGa_imperv,VGn_imperv:
impervious surface hydraulic conductivity (m/hr), porosity (-), residual saturation
(-), saturation (-), van Genuchten alpha (1/m), and van Genuchten n (-).
Ks_soil,porosity_soil,Sres_soil,Ssat_soil,VGa_soil,VGn_soil: as above, for soil.
mn_grass,mn_imperv: Manning's n (hr*m1/3) for turfgrass and impervious surfaces
parcelCover,fc: parcel cover indicating feature cover and coordinates of the
impervious features (see data/layouts) NaNimp,pervX,pervY: matrix with NaNs at
impervious pixels, for viewing results as well as row and column of random pervious
element (sometimes helpful to have during post-processing visualization)
slopeX,slopeY: matrices with slope in x and slope in y directions
data/soil
Includes subsurface hydraulic parameters for impervious surfaces (imperv), silt
loam (SiL), moderately-compacted silt loam (SiL2c), and highly-compacted silt loam
(SiL10c) baseline soil conditions. This information is ultimately incorporated into
domainInfo.mat (see data/model_inputs).
data/weather
Includes meterological forcing information for average and dry growing season
weather scenarios, as well as non-changing CLM inputs. Each subdirectory includes
the following files:
drv_clmin_start.dat,drv_clmin_restart.dat: CLM timing information for new start
and restart models.
drv_vegp.dat: CLM vegetation parameters (LAI, rooting parameters, etc.) for
each landcover type.
nldas.1hr.clm.txt: hourly meteorological inputs needed for CLM. Columns are 1)
DSWR, shortwave radiation (W/mless than sup.2), 2)
DLWR, longwave radiation (W/mless than sup.2), 3) APCP, precipitation (mm/s),
4) Temp, air temperature (K), 5) UGRD, east-west wind speed (m/s), 6) VGRD,
north-south wind speed (m/s), 7) Press, atmospheric pressure (pa), 8) SPFH, specific
humidity (kg/kg).
precip.mat: hourly precipitation timeseries (m) extracted from nldas.1hr.clm.txt
for use in Matlab post-processing.
results
Due to size of output files, only a limited selection of output files are
included in this repo (i.e., those used to create manuscript figures and hourly parcel
fluxes).
Model subdirectories are named acording to the same convention in
data/model_inputs. Each subdirectory may include the following files:
WBstep.mat: suite of variables with the hourly flux (m3) at each hour for all
hydrologic fluxes (can = water stored in the canopy, dd = deep drainge, etS =
evaptranssum, ev = evaporation, precip = precipitation, re = recharge, sno = snow
water equivalent, sr = surface runoff, Ss = surface storage, Sss = subsurface storage,
SssRZ = subsurface storage in the root zone aka top 1m). Files also includes the
hourly forcing (force) for each model component (CLM, PF = parflow, O = overall), the
hourly ouputs and storage (calc), and the difference between the forcing and
calculated fluxes as a volume (absErr) and relative to the forcing (relErr).
WBtotal.mat (developed lots) or WBcum.mat (vacant lots): as with WBstep.mat, but
with the running cumulative flux at each hour for all hydrologic fluxes. Due to
changes in post-processing scripts, fluxes for developed lots are as a volume (m3),
while fluxes for vacant lots are as a depth (mm).
deep_drainage.grid.cum.mat: matrix (nx X ny) with the cumulative growing season
deep drainage (m3) for each model element.
evaporation.grid.cum.mat: as deep_drainage.grid.cum.mat, but for evaporation (from
leaves and soil).
transpiration.grid.cum.mat: as deep_drainage.grid.cum.mat, but for transpiration.
src
src/manuscript_figures
Scripts here create manuscript figures based on files in
results/selected_model_outputs. Files include:
figure02_example_layout.m: generates layout of lowest-impact parcel (all 5
interventions applied) with elevation indicated via heatmap.
figures04_05_06_diffs_pairs.m: calculates difference between each simulation and
the baseline scenario as a depth and as a percent, as displayed in figures 4 and 5.
Also calculates the synergistic effects of combining low impact interventiosn, as
displayed in figure 6.
figure07_compare_weather.m: compares cumulative growing season deep drainage
and evapotranspiration as spatial maps for the lowest impact lot (all 5
interventions applied) under average and dry weather scenarios.
figure08_compare_vacant_lot.m: compares total growing season runoff, deep
drainage, evapotranspiration, and transpiration per unit vegetated area for the
highly-compacted baseline, lowest-impact lot, and vacant lot under average and dry
weather scenarios.
src/model_inputs
Scripts here create input directories for all 96 model simulations (in
data/model_inputs) based on files in 'data' directory. Files include:
lot_microtopography.m: demonstrates how deviations in elevation were randomly
generated for microtopography scenarios.
lot_layouts.m: uses lot data in data/layouts to calculate slopes and generate
.sa input files for parflow based on downspout, sidewalk, transverse slope, and
microtopography features
lot_slopes.m: fills in pits in microtopography elevation (if appropriate),
calculates slopes on pervious pixels, then defines slopes for impervious
features.
matrixTOpfsa.m: converts Matlab matrix (nx X ny) into .sa file suitable for
input to parflow as pfsa file type.
matrixTOvegm.m: converts Matlab matrix (nx X ny) of vegetation land cover type
into CLM input file drv_vegm.dat.
model_inputs.m: uses lot data generated by lot_layouts.m and adds information
about soils and weather based on model scenario. After running this script, complete
set of input files for all developed models resides in data/model_inputs.
src/model_outputs
Scripts here copy and evaluate model outputs. Files include:
copy_model_output.m: script used to copy selected model output (hourly fluxes)
from master directory with all parflow model outputs to this repository.
assess_model_errors.m. extracts overall absolute (mm) and relative (-) model
error for each model and plots for visualization. "Error" here is the difference
between model forcing (precipitation) and all other fluxes (outflows + change in
storage). Relative error is calculcated relative to model forcing
(precipitation).
src/runParflow.tcl
Version of parflow executable used to run these models. Note that for versions
of parflow from 05/01/2017 to present (6/11/2018), there is a bug in how parflow
interprets the Patch.bottom.BCPressure Dirichlet boundary condition. Related
repositories for running parflow:
cvoter/parflow: forked from parflow/parflow.
cvoter/PFinstall: notes and scripts for installing parflow
cvoter/PFscripts: wrapper scripts for this executable, including additonal pre-
and post-processing scripts.
src/plot_turfgrass_roots.m
Matlab script compares rooting parameters used in Parflow.CLM model to
empirical observations of turgrass rooting depth from the literature.
