Post on 07-Aug-2020
Geophysical Fluid Dynamics Laboratory Review
May 20 – May 22, 2014
Presented by
Hydrology
Chris Milly
Geophysical Fluid Dynamics Laboratory Review
May 20-22, 2014
Hydrology: Preview
①Land model LM3, used in all GFDL CMIP5
streams, contains new physical/hydrologic
features that
• improve fluxes to atmosphere and oceans;
• support terrestrial biogeochemical modeling;
• represent impact-relevant hydrologic variables (streamflow, lake levels, water table).
②A focus on potential evapotranspiration (PET)
helps reconcile conflicting projections of
water availability and drought.
1
Geophysical Fluid Dynamics Laboratory Review
May 20-22, 2014
LM3 Hydrology: Features
2
soil-veg column (tile)
lake tile
landscape unit
grid cell
river network
Geophysical Fluid Dynamics Laboratory Review
May 20-22, 2014
lake
leve
ls
LM3 Hydrology: Stand-Alone Evaluation
3
annual range of water storage (mm)
LM3 storage variables compare well with observations.
Geophysical Fluid Dynamics Laboratory Review
May 20-22, 2014
LM3 Hydrology: Evaluation
4
permafrost extent
late
nt
hea
t fl
ux
Observations
LM3
DICE: Diurnal Land/Atmosphere Coupling Experiment
(credit: Adrian Lock, Martin Best)
LM3 reproduces permafrost extent well.
Geophysical Fluid Dynamics Laboratory Review
May 20-22, 2014
Work in Progress: Toward LM4
5
Hill-slope tiling (Subin) Urban tiling
+ Higher Resolution + Biogeochemistry (Shevliakova)
Geophysical Fluid Dynamics Laboratory Review
May 20-22, 2014
Potential Evapotranspiration
6
Climate
Projections
Empirical Relations
(e.g., Thornthwaite)
Hydrologic Models,
Drought Metrics
(e.g., PDSI)
P
T PET Water
Projections
Many hydrologic impact analyses have PET as a
pivotal link:
Geophysical Fluid Dynamics Laboratory Review
May 20-22, 2014
Potential Evapotranspiration
7
• In a case study, PET change
had been overestimated by a
factor of ~3 vs. that implicit in
the climate model.
• Substantial negative bias in
runoff change results.
• SREXAR5: “AR4 conclusions
regarding global increasing
drought since the 1970s were
probably overstated.”
y = 3.0x
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 0.2
Relative Change in PET
climate model
hyd
rolo
gic
mo
del
(after Milly and Dunne [2011])
Geophysical Fluid Dynamics Laboratory Review
May 20-22, 2014
Hydrology: Recap
①Land model LM3, used in all GFDL CMIP5
streams, contains new physical/hydrologic
features that
• improve fluxes to atmosphere and oceans;
• support terrestrial biogeochemical modeling;
• represent impact-relevant hydrologic variables (streamflow, lake levels, water table).
②A focus on potential evapotranspiration (PET)
helps reconcile conflicting projections of
water availability and drought.
8