Representing nighttime and minimum conductance in CLM4.5: global hydrology and carbon sensitivity analysis using  observational constraints

Lombardozzi, Danica L.; Zeppel, Melanie J. B.; Fisher, Rosie A.; Tawfik, Ahmed

doi:https://doi.org/10.5194/gmd-10-321-2017

Articles | Volume 10, issue 1

https://doi.org/10.5194/gmd-10-321-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/gmd-10-321-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 10, issue 1

Model experiment description paper

| Highlight paper

|

23 Jan 2017

Model experiment description paper | Highlight paper |

| 23 Jan 2017

Representing nighttime and minimum conductance in CLM4.5: global hydrology and carbon sensitivity analysis using observational constraints

Danica L. Lombardozzi, Melanie J. B. Zeppel, Rosie A. Fisher, and Ahmed Tawfik

Download

Final revised paper (published on 23 Jan 2017)
Supplement to the final revised paper
Preprint (discussion started on 04 Dec 2015)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

SC C3222: 'Executive Editor Comment on "Observed nighttime conductance alters modeled global hydrology and carbon budgets"', Astrid Kerkweg, 07 Dec 2015
SC C3521: 'Review of Lombardozzi et al. (in review) Observed nighttime conductance alters modeled global hydrology and carbon budgets', JOSHUA FISHER, 29 Dec 2015
- RC C3561: 'Referee Comment', J. B. Fisher, 05 Jan 2016
RC C3562: 'Review of "Observed nighttime conductance alters modeled global hydrology and carbon budgets"', Anonymous Referee #2, 05 Jan 2016
AC C4063: 'Author Comments and Manuscript Revisions', Danica Lombardozzi, 26 Feb 2016

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Danica Lombardozzi on behalf of the Authors (15 Mar 2016) Author's response Manuscript

ED: Referee Nomination & Report Request started (10 May 2016) by Tomomichi Kato

RR by Kevin Tu (12 Aug 2016)

Suggestions for revision or reasons for rejection

Consistent with the first two reviewers, in particular reviewer
Fisher, I
believe the motivation
behind this study is excellent, but that there are some issues that
still
need to be addressed
which I describe below.

On reviewer Fisher's suggestion for comparison to empirical
observations,
this study still lacks
sufficient validation, for example the data presented in Figure 5 is
poor
at best for validating
the model. Without any means to gauge if the model predictions are
reasonable, this study is most
effectively a sensitivity analysis that serves to highlight the need
for
additional studies,
particularly field measurements, on nighttime conductance and how it
is
parameterized in land
surface models.

I agree with the 2nd reviewer's comments on the misleading title and
attribution of the effects
to nighttime conductance not being accurate - that the results
indicate
larger impacts on water
and carbon fluxes occur when the minimum conductance during the day
is
altered rather than the
nighttime conductance. Further, the authors agree and point out that
the
focus is really on
stomatal conductance rather than nighttime or daytime minimum
conductance
per se, and as such,
the title should not focus solely on minimum conductance. However,
by the
author's own logic,
the title should not focus solely on nighttime conductance either.

The authors also argue that it's unclear whether the observations of
nighttime conductance are
equivalent to minimum conductance, thus their emphasis on nighttime
conductance in the title.
While the authors may have used observations of nighttime
conductance,
they clearly used them to
alter both nighttime and minimum daytime conductance in the model
(in the
?g0 and ?gmin
scenarios), with the result being larger effects of minimum daytime
conductance. Both the methods
and the results are not consistent with the bias towards nighttime
conductance in the title.

Further, not only were observations of nighttime conductance
incorporated
into the model, but a
new parameterization of minimum daytime conductance was effectively
incorporated as well (in the
?gmin scenario). Given that the title should reflect the content of
the
paper, the title should
indicate to both nighttime and minimum conductance. This may not
have
been the original objective
of the authors, but it does reflect what they actual did. In two of
three
approaches examined
(the ?g0 and ?gmin scenarios), the conductances were altered during
both the nighttime and
daytime.

The authors indicate (L156) that the observations are most likely
representative of maximum
nighttime conductance, since most measurements were done during
well-watered conditions without
water stress, yet inconsistent with this, the observations were
incorporated in the first
approach (?g0) as actual nighttime conductance (no soil moisture
constraint was used).
Substituting observations of maximum nighttime conductance directly
for
values of actual
nighttime conductance seems illogical and physiologically unrealistic.

The real value and contribution of this study is as a sensitivity
analysis of the BWB model,
specifically the sensitivity of global water and carbon fluxes to
the
value of the BWB intercept
in the CLM4.5. This would be consistent with the author's statement
in
response to reviewer
Fisher's comment on the need for more empirical validation, that the
"…
primary aim is to
highlight the high sensitivity of the hydrological and carbon cycles
to
these typically poorly
constrained parameters". The physical meaning of the BWB intercept
and/or its physiological
interpretation is debatable. Regardless, it's important to
characterize
and understand the
sensitivity of the CLM4.5, and water and carbon budgets in general,
to
this variable. As noted by
reviewer Fisher, nighttime conductance and associated nighttime
transpiration is an
under-represented but potentially important process in models of
land-atmosphere interaction.
Clearly, this is what the authors set out to address. However, their
efforts are confounded by
the fact that the current parameterization of stomatal conductance
using
the BWB model in the
CLM4.5 is not easily modified to include nighttime conductance. The
issues raised by the
reviewers highlight the empirical side of the BWB model, and the
problems
associated with
attempting a mechanistic implementation (i.e. nighttime conductance)
of a
largely empirical model
(i.e. the BWB intercept).

Further, the observations are too far and few between to be
representative of the actual BWB
intercept (night or day) for a given PFT in a global model like CLM4.5.
It would make more sense
to use the observations merely as realistic constraints on the range
of
potential variation of
the BWB intercept in a sensitivity analysis. Further still, if the
objective is to determine the
sensitivity of CLM4.5 to nighttime conductance it should be
sufficient to
examine only ?gnight.
Once changes to daytime conductance are made (e.g. through changes
in
minimum conductance) the
question then expands to daytime as well as nighttime conductance,
which
is really beyond the
goal of this study. If the goal is truly focused on nighttime
conductance
then simply change
nighttime conductance, and exclude both the ?gmin experiment which
involves changes to daytime
conductance as well as the ?g0 experiment, which also includes
changes
to daytime conductance
(and is unreasonable for other reasons as well, see above). If the
goal
is really to highlight
the sensitivity of the model to nighttime conductance then the issue
of
daytime conductance being
consistent with the nighttime conductance is beyond the scope of the
study. More than anything,
the ?gmin experiment should be included only for discussion
purposes, to
address the issue of
consistency between nighttime and daytime conductance, rather than
as an
alternative method of
modifying nighttime conductance. The authors could then focus the
manuscript on their stated goal
of "Sensitivity of global water and carbon budgets to nighttime
conductance in CLM4.5".

On reviewer Fisher's suggestion to incorporate the results
synthesized
in a Tree Physiology
special issue in 2007, I do not feel the authors responded adequately.
First, in contrast to the
author's claim that the papers in that special issue do not include
environmental sensitivities
of nighttime conductance, the paper by Dawson et al. provides a
clear
relationship between
observations of nighttime conductance and days following rainfall,
and
changes in the ratio of
nighttime to daytime conductance following rainfall, with greater
fractions (~25%) during the
wettest periods immediately following rainfall, with a decline of 5%
per
day after rain. As
noted, "This relationship provides a strong and predictable index of
water loss from plants at
night based on daylight values…". This type of data should provide
valuable information for
parameterizing nighttime conductance as a function of daytime
conductance
and time since rainfall
or soil moisture. Second, the authors note that some plant types are
sensitive to environmental
factors while others are not. This phenomena needs to be explained
rather
than used as evidence
to discard the data. It could very well be that nighttime
conductance is
not a phenomenon
parameterized as easily as the intercept of the BWB model
constrained by
soil moisture.

Specific Comments

Equation 1: The full equation including the soil adjustment factor
(?
soil) should be shown. Not
showing the full equation can be confusing and potentially
misleading. It
would be informative to
also show the ?soil parameterization and the parameter values by
PFT,
since poor parametrization
of this function could lead to poor performance and unreasonable
results,
for example if ?soil
did not adequately constrain g0 during drought conditions (e.g. in
semi-arid regions).

It's worth noting that in CLM4.5, soil drought also effectively
impacts
g1 by way of
constraining Vcmax (I'm assuming this based on the fact that the
authors
state CLM is based on
SiB2; Sellers et al. 1996).

Are there differences between glasshouse and field gs,n values?
Using
glasshouse gs,n data needs
to be validated since glasshouse conditions, including both plant
and
environmental, can be
unrepresentative of actual field conditions and plant responses in
the
field. In the very least,
glasshouse vs field data should be clearly indicated in Table 1.

L115: It's not clear what the 'simulated PFT' is. Were the PFTs
'simulated' then replaced
with observed values of g0? It probably just needs wordsmithing. If
so,
it seems clearer to
simply state that constant minimum gs values were assumed for each
PFT
for the method.

References are needed in Table 1 to know the source of the data.

The data in Table 1 indicates that the entirety of boreal forests,
both
needleleaf evergreen and
broadleaf deciduous, are each represented by a single measurement.
This
is poor parameter
estimation at best and reinforces the notion that the observations
would
serve best as a guide
for their potential range in a sensitivity analysis, rather than as
direct estimates of nighttime
conductance. Given the paucity of data, its unreasonable to expect
that
the observations will be
robust representations of actual values of (maximum) nighttime
conductance for all plants at
every time-step throughout every growing season within each PFT.

Values in Table 1 should only be reported to significant digits.
It's
uninformative and
potentially misleading to indicate conductance out to 8 decimal places.

Hide

ED: Reconsider after major revisions (07 Sep 2016) by Tomomichi Kato

AR by Danica Lombardozzi on behalf of the Authors (18 Oct 2016)

ED: Referee Nomination & Report Request started (07 Nov 2016) by Tomomichi Kato

RR by Kevin Tu (12 Dec 2016)

ED: Publish as is (19 Dec 2016) by Tomomichi Kato

AR by Danica Lombardozzi on behalf of the Authors (27 Dec 2016)

Short summary

Earth's terrestrial surface influences climate by exchanging carbon and water with the atmosphere through stomatal pores. However, most land-surface models, used to predict global carbon and water fluxes, estimate that water lost through stomata is less than what observations show. In this study, we integrate plant water loss data from 204 species into a global land surface model, finding that global estimates of plant water loss increase, soil moisture decreases, and carbon gain also decreases.