2.20. Vegetation Phenology and Turnover¶
The CLM phenology model consists of several algorithms controlling the transfer of stored carbon and nitrogen out of storage pools for the display of new growth and into litter pools for losses of displayed growth. PFTs are classified into three distinct phenological types that are represented by separate algorithms: an evergreen type, for which some fraction of annual leaf growth persists in the displayed pool for longer than one year; a seasonal-deciduous type with a single growing season per year, controlled mainly by temperature and daylength; and a stress-deciduous type with the potential for multiple growing seasons per year, controlled by temperature and soil moisture conditions.
The three phenology types share a common set of control variables. The calculation of the phenology fluxes is generalized, operating identically for all three phenology types, given a specification of the common control variables. The following sections describe first the general flux parameterization, followed by the algorithms for setting the control parameters for the three phenology types.
2.20.1. General Phenology Flux Parameterization¶
Fluxes of carbon and nitrogen from storage pools and into displayed tissue pools pass through a special transfer pool (denoted _xfer), maintained as a separate state variable for each tissue type. Storage (_stor) and transfer (_xfer) pools are maintained separately to reduce the complexity of accounting for transfers into and out of storage over the course of a single growing season.
![../../_images/image110.png](../../_images/image110.png)
Figure 2.20.1 Example of annual phenology cycle for seasonal deciduous.¶
2.20.1.1. 14.1.1 Onset Periods¶
The deciduous phenology algorithms specify the occurrence of onset growth periods (Figure 14.1). Carbon fluxes from the transfer pools into displayed growth are calculated during these periods as:
(2.20.1)¶
(2.20.2)¶
(2.20.3)¶
(2.20.4)¶
(2.20.5)¶
(2.20.6)¶
with corresponding nitrogen fluxes:
(2.20.7)¶
(2.20.8)¶
(2.20.9)¶
(2.20.10)¶
(2.20.11)¶
(2.20.12)¶
where CF is the carbon flux, CS is stored carbon, NF is the nitrogen
flux, NS is stored nitrogen, (s-1) is a time-varying rate coefficient controlling flux
out of the transfer pool:
(2.20.13)¶
and tonset (s) is the number of seconds remaining in the current phenology onset growth period (Figure 14.1). The form of Eq. (2.20.13) produces a flux from the transfer pool which declines linearly over the onset growth period, approaching zero flux in the final timestep.
2.20.1.2. 14.1.2 Offset Periods¶
The deciduous phenology algorithms also specify the occurrence of litterfall during offset periods. In contrast to the onset periods, only leaf and fine root state variables are subject to litterfall fluxes. Carbon fluxes from display pools into litter are calculated during these periods as:
(2.20.14)¶
(2.20.15)¶
(2.20.16)¶
where superscripts n and n-1 refer to fluxes on the current and
previous timesteps, respectively. The rate coefficient varies with time to produce a linearly
increasing litterfall rate throughout the offset period, and the special
case for fluxes in the final litterfall timestep
(
=
) ensures that all of the
displayed growth is sent to the litter pools for deciduous plant types.
Corresponding nitrogen fluxes during litterfall take into account retranslocation of nitrogen out of the displayed leaf pool prior to
litterfall (, gN m-2 s-1). Retranslocation of nitrogen out of fine roots is
assumed to be negligible. The fluxes are:
(2.20.17)¶
(2.20.18)¶
(2.20.19)¶
where CN is C:N.
2.20.1.3. 14.1.3 Background Onset Growth¶
The stress-deciduous phenology algorithm includes a provision for the
case when stress signals are absent, and the vegetation shifts from a
deciduous habit to an evergreen habit, until the next occurrence of an
offset stress trigger . In that case, the regular onset flux mechanism
is switched off and a background onset growth algorithm is invoked
(). During this period, small fluxes
of carbon and nitrogen from the storage pools into the associated
transfer pools are calculated on each time step, and the entire contents
of the transfer pool are added to the associated displayed growth pool
on each time step. The carbon fluxes from transfer to display pools
under these conditions are:
(2.20.20)¶
(2.20.21)¶
(2.20.22)¶
(2.20.23)¶
(2.20.24)¶
(2.20.25)¶
and the corresponding nitrogen fluxes are:
(2.20.26)¶
(2.20.27)¶
(2.20.28)¶
(2.20.29)¶
(2.20.30)¶
(2.20.31)¶
2.20.1.4. 14.1.4 Background Litterfall¶
Both evergreen and stress-deciduous phenology algorithms can specify a
litterfall flux that is not associated with a specific offset period,
but which occurs instead at a slow rate over an extended period of time,
referred to as background litterfall. For evergreen types the background
litterfall is the only litterfall flux. For stress-deciduous types
either the offset period litterfall or the background litterfall
mechanism may be active, but not both at once. Given a specification of
the background litterfall rate (, s-1), litterfall carbon fluxes are calculated as
(2.20.32)¶
(2.20.33)¶
with corresponding nitrogen litterfall and retranslocation fluxes:
(2.20.34)¶
(2.20.35)¶
(2.20.36)¶
2.20.1.5. 14.1.5 Livewood Turnover¶
The conceptualization of live wood vs. dead wood fractions for stem and coarse root pools is intended to capture the difference in maintenance respiration rates between these two physiologically distinct tissue types. Unlike displayed pools for leaf and fine root, which are lost to litterfall, live wood cells reaching the end of their lifespan are retained as a part of the dead woody structure of stems and coarse roots. A mechanism is therefore included in the phenology routine to effect the transfer of live wood to dead wood pools, which also takes into account the different nitrogen concentrations typical of these tissue types.
A live wood turnover rate (, s-1) is
defined as
(2.20.37)¶
where is the assumed annual live wood
turnover fraction. Carbon fluxes from live to dead wood pools are:
(2.20.38)¶
(2.20.39)¶
and the associated nitrogen fluxes, including retranslocation of nitrogen out of live wood during turnover, are:
(2.20.40)¶
(2.20.41)¶
(2.20.42)¶
(2.20.43)¶
2.20.2. Evergreen Phenology¶
The evergreen phenology algorithm is by far the simplest of the three
possible types. It is assumed for all evergreen types that all carbon
and nitrogen allocated for new growth in the current timestep goes
immediately to the displayed growth pools (i.e. f
(Chapter 13)). As such, there is never an accumulation of carbon or
nitrogen in the storage or transfer pools, and so the onset growth and
background onset growth mechanisms are never invoked for this type.
Litterfall is specified to occur only through the background litterfall
mechanism – there are no distinct periods of litterfall for evergreen
types, but rather a continuous (slow) shedding of foliage and fine
roots. This is an obvious area for potential improvements in the model,
since it is known, at least for evergreen needleleaf trees in the
temperate and boreal zones, that there are distinct periods of higher
and lower leaf litterfall (Ferrari, 1999; Gholz et al., 1985). The rate
of background litterfall (
, section 14.1.4)
depends on the specified leaf longevity (
, y), as
(2.20.44)¶
2.20.3. Seasonal-Deciduous Phenology¶
The seasonal-deciduous phenology algorithm derives directly from the
treatment used in the offline model Biome-BGC v. 4.1.2, (Thornton et
al., 2002), which in turn is based on the parameterizations for leaf
onset and offset for temperate deciduous broadleaf forest from White et
al. (1997). Initiation of leaf onset is triggered when a common
degree-day summation exceeds a critical value, and leaf litterfall is
initiated when daylength is shorter than a critical value. Because of
the dependence on daylength, the seasonal deciduous phenology algorithm
is only valid for latitudes outside of the tropical zone, defined here
as . Neither the
background onset nor background litterfall mechanism is invoked for the
seasonal-deciduous phenology algorithm. The algorithm allows a maximum
of one onset period and one offset period each year.
The algorithms for initiation of onset and offset periods use the winter
and summer solstices as coordination signals. The period between winter
and summer solstice is identified as ,
and the period between summer and winter
solstice is identified as
,
where
and
are the day length(s) calculated for the
current and previous timesteps, respectively, using
(2.20.45)¶
where lat and decl are the latitude and solar declination (radians), respectively, and the factor 13750.9871 is the number of seconds per radian of hour-angle.
2.20.3.1. 14.3.1 Seasonal-Deciduous Onset Trigger¶
The onset trigger for the seasonal-deciduous phenology algorithm is
based on an accumulated growing-degree-day approach (White et al.,
1997). The growing-degree-day summation () is
initiated (
) when the phenological state is
dormant and the model timestep crosses the winter solstice. Once these
conditions are met,
is updated on each timestep as
(2.20.46)¶
where (K) is the temperature of the third soil layer, and
.
The onset period is initiated if
,
where
(2.20.47)¶
and where (K) is the annual average of
the 2m air temperature, and TKFRZ is the freezing point of water (273.15 K). The following control variables are set when a new onset growth
period is initiated:
(2.20.48)¶
(2.20.49)¶
where is set to a constant value of 30 days.
Fluxes from storage into transfer pools occur in the timestep when a new
onset growth period is initiated. Carbon fluxes are:
(2.20.50)¶
(2.20.51)¶
(2.20.52)¶
(2.20.53)¶
(2.20.54)¶
(2.20.55)¶
(2.20.56)¶
and the associated nitrogen fluxes are:
(2.20.57)¶
(2.20.58)¶
(2.20.59)¶
(2.20.60)¶
(2.20.61)¶
(2.20.62)¶
where is the fraction of current storage
pool moved into the transfer pool for display over the incipient onset
period. This fraction is set to 0.5, based on the observation that
seasonal deciduous trees are capable of replacing their canopies from
storage reserves in the event of a severe early-season disturbance such
as frost damage or defoliation due to insect herbivory.
If the onset criterion () is not met before the summer solstice,
then
is set to 0.0 and the growing-degree-day
accumulation will not start again until the following winter solstice.
This mechanism prevents the initiation of very short growing seasons
late in the summer in cold climates. The onset counter is decremented on
each time step after initiation of the onset period, until it reaches
zero, signaling the end of the onset period:
(2.20.63)¶
2.20.3.2. 14.3.2 Seasonal-Deciduous Offset Trigger¶
After the completion of an onset period, and once past the summer
solstice, the offset (litterfall) period is triggered when daylength is
shorter than 39300 s. The offset counter is set at the initiation of the
offset period: , where
is set to a constant value of 15 days. The
offset counter is decremented on each time step after initiation of the
offset period, until it reaches zero, signaling the end of the offset
period:
(2.20.64)¶
2.20.4. Stress-Deciduous Phenology¶
The stress-deciduous phenology algorithm was developed specifically for the CLM based in part on the grass phenology model proposed by White et al. (1997). The algorithm handles phenology for vegetation types such as grasses and tropical drought-deciduous trees that respond to both cold and drought-stress signals, and that can have multiple growing seasons per year. The algorithm also allows for the possibility that leaves might persist year-round in the absence of a suitable stress trigger. In that case the phenology switches to an evergreen habit, maintaining a marginally-deciduous leaf longevity (one year) until the occurrence of the next stress trigger.
2.20.4.1. 14.4.1 Stress-Deciduous Onset Triggers¶
In climates that are warm year-round, onset triggering depends on soil
water availability. At the beginning of a dormant period (end of
previous offset period), an accumulated soil water index
(, d) is initialized (
), with subsequent accumulation calculated as:
(2.20.65)¶
where s,3 is the soil water potential (MPa)
in the third soil layer and
is the onset soil water potential threshold. Onset triggering is
possible once
. To avoid spurious onset triggering due to
soil moisture in the third soil layer exceeding the threshold due only to
soil water suction of water from deeper in the soil column, an additional precipitation trigger is included which requires
at least 20 mm of rain over the previous 10 days (Dahlin et al., 2015). If the cold climate
growing degree-day accumulator is not active at the time when the soil moisture and precipitation
thresholds are reached (see below), and if the daylength is greater than 6
hours, then onset is triggered. Except as noted below,
continues to accumulate according to Eq. (2.20.65) during
the dormant period if the daylength criterion prevents onset triggering,
and onset is then triggered at the timestep when daylength exceeds 6
hours.
In climates with a cold season, onset triggering depends on both
accumulated soil temperature summation and adequate soil moisture. At
the beginning of a dormant period a freezing day accumulator
(, d) is initialized (
),
with subsequent accumulation calculated as:
(2.20.66)¶
If during the dormant period, then a
cold-climate onset triggering criterion is introduced, following exactly
the growing degree-day summation (
) logic of Eqs. (2.20.46)
and (2.20.47). At that time
is reset
(
). Onset triggering under these conditions
depends on meeting all three of the following criteria:
,
, and daylength greater than 6 hrs.
The following control variables are set when a new onset growth period
is initiated: ,
,
,
, and
, where
is set to a constant value of 30 days. Fluxes
from storage into transfer pools occur in the timestep when a new onset growth period is initiated, and are handled identically to Eqs. (2.20.50) -(2.20.56) for
carbon fluxes, and to Eqs. (2.20.57) - (2.20.62) for nitrogen fluxes. The onset counter is decremented on each time step after initiation of the onset period,
until it reaches zero, signaling the end of the onset period:
(2.20.67)¶
2.20.4.2. 14.4.2 Stress-Deciduous Offset Triggers¶
Any one of the following three conditions is sufficient to initiate an
offset period for the stress-deciduous phenology algorithm: sustained
period of dry soil, sustained period of cold temperature, or daylength
shorter than 6 hours. Offset triggering due to dry soil or cold
temperature conditions is only allowed once the most recent onset period
is complete. Dry soil condition is evaluated with an offset soil water
index accumulator (, d). To test for a sustained
period of dry soils, this control variable can increase or decrease, as
follows:
(2.20.68)¶
where is the offset soil
water potential threshold. An offset period is triggered if the previous
onset period is complete and
, where
.
The cold temperature trigger is calculated with an offset freezing day
accumulator (, d). To test for a sustained period
of cold temperature, this variable can increase or decrease, as follows:
(2.20.69)¶
An offset period is triggered if the previous onset period is complete
and ,
where
.
The offset counter is set at the initiation of the offset period:
, where
is set to a constant value of 15 days. The
offset counter is decremented on each time step after initiation of the
offset period, until it reaches zero, signaling the end of the offset
period:
(2.20.70)¶
2.20.4.3. 14.4.3 Stress-Deciduous: Long Growing Season¶
Under conditions when the stress-deciduous conditions triggering offset
are not met for one year or longer, the stress-deciduous algorithm
shifts toward the evergreen behavior. This can happen in cases where a
stress-deciduous vegetation type is assigned in a climate where suitably
strong stresses occur less frequently than once per year. This condition
is evaluated by tracking the number of days since the beginning of the
most recent onset period (, d). At the end
of an offset period
is reset to 0. A long
growing season control variable (LGS, range 0 to 1) is calculated as:
(2.20.71)¶
The rate coefficient for background litterfall (, s-1) is calculated as a function of LGS:
(2.20.72)¶
where is the leaf longevity. The result is a shift to continuous litterfall as
increases from 365 to 730. When a new offset period is triggered
is set to 0.
The rate coefficient for background onset growth from the transfer pools ( , s-1) also depends on LGS, as:
(2.20.73)¶
On each timestep with
0, carbon fluxes from storage to transfer pools are calculated as:
(2.20.74)¶
(2.20.75)¶
(2.20.76)¶
(2.20.77)¶
(2.20.78)¶
(2.20.79)¶
with corresponding nitrogen fluxes:
(2.20.80)¶
(2.20.81)¶
(2.20.82)¶
(2.20.83)¶
(2.20.84)¶
(2.20.85)¶
The result, in conjunction with the treatment of background onset
growth, is a shift to continuous transfer from storage to display pools
at a rate that would result in complete turnover of the storage pools in
one year at steady state, once LGS reaches 1 (i.e. after two years
without stress-deciduous offset conditions). If and when conditions
cause stress-deciduous triggering again, is rest
to 0.
2.20.5. Litterfall Fluxes Merged to the Column Level¶
CLM uses three litter pools, defined on the basis of commonly measured chemical fractionation of fresh litter into labile (LIT1 = hot water and alcohol soluble fraction), cellulose/hemicellulose (LIT2 = acid soluble fraction) and remaining material, referred to here for convenience as lignin (LIT3 = acid insoluble fraction) (Aber et al., 1990; Taylor et al., 1989). While multiple plant functional types can coexist on a single CLM soil column, each soil column includes a single instance of the litter pools. Fluxes entering the litter pools due to litterfall are calculated using a weighted average of the fluxes originating at the PFT level. Carbon fluxes are calculated as:
(2.20.86)¶
(2.20.87)¶
(2.20.88)¶
(2.20.89)¶
(2.20.90)¶
(2.20.91)¶
where ,
, and
are the labile, cellulose/hemicellulose,
and lignin fractions of leaf litter for PFT p,
,
, and
are the labile, cellulose/hemicellulose,
and lignin fractions of fine root litter for PFT p,
is the weight relative to the column for PFT
p, and p is an index through the plant functional types occurring on
a column. Nitrogen fluxes to the litter pools are assumed to follow the
C:N of the senescent tissue, and so are distributed using the same
fractions used for carbon fluxes:
(2.20.92)¶
(2.20.93)¶
(2.20.94)¶
(2.20.95)¶
(2.20.96)¶
(2.20.97)¶