IPILPS Workshop

1
IsoTransIsotopes in the boundary
layerAlastair Williams

• IPILPS Workshop
• ANSTO 18-22 April 2005

2
Introduction

• IsoTrans (Isotopic Tracers in Atmospheric
  Transport)
• ANSTO mother Project for IPILPS
• Broader scope
• Purpose of presentation
• Introduction to IsoTrans (very short)
• How is IsoTrans contributing to the improvement
  of surface and boundary layer representations in
  models?
• Evaluation and development of LSSs in
  isotope-enabled hydroclimate models (main subject
  of current workshop)
• How can SWI obs add value to our understanding
  of moisture exchange in plant canopies,
  particularly ET partitioning?
• Natural radionuclides and turbulent mixing in the
  lower atmosphere

3
IsoTrans Drivers
Effective Environmental Management Strategies
need Informed predictions of mixing and movement
4
IsoTrans Drivers
Effective Environmental Management Strategies
need Informed predictions of mixing and movement
5
IsoTrans 3 Foci, 3 Scales
6
IsoTrans Process Studies

• IsoTrans Task 3 (IPILPS)
• How can SWI observations add value to our
  understanding of moisture exchange in plant
  canopies, particularly evapo-transpiration
  partitioning?
• Discuss the Keeling approach for estimating the
  transpired component of ET in vegetation canopies
• Examine turbulent transport within vegetation
  canopies
• Analyze SWI behaviour in Tumbarumba air space
• Present first guess at ET partition for
  Tumbarumba
• Thanks to David Griffith (Wollongong Uni) for
  providing the vertical ?D data, and Helen Cleugh
  / Ray Leuning for providing the met data

7
Use of SWIs to Partition ET
Concept simple mix of 2 fluxes with distinct
isotopic signatures (?) evap (frac) and transp
(non-frac)
?T, ?E composition of contributing sources
(measured / calculated) ?ET effective
combined source FET from EC ? FT How to
estimate ?ET?
8
Keeling (1958)

• Carbon isotope ratio closely follows concn in
  diurnal time series over different vegetated
  surfaces
• Mutual variation suggests simple 2-part mixing

(air and plants)
9
Keeling Analysis (1)
2-part mixing model (ambient combined ET)
Cm, ?mx measured Ca, ?ax background component
from atmosphere CET, ?ETx combined component
from evap and transp
? Linear relation if Ca, ?ax and ?ETx constant,
with intercept ?ETx
10
Keeling Analysis (2)

• Versatile (temporal vertical gradients)
• Problems
• Extrapolated intercept susceptible to large
  errors
• Breakdown of assumptions

1. Simple mixing of two major sources/sinks (atmos
   ET)
2. No sources/sinks other than evap transp (eg.
   dew, fog)
3. Relative contribution of all subsources remains
   fixed (eg. non-fractionating
   transpiration assumption true only when averaged
   over whole day Harwood et al. 1998)

11
Harwood et al. (1998)
Diurnal variation of 18O of transpired water
vapour for leaves on day 1 (?) and day 2 ( ?,?,?)
indicating the vapour pressure deficit (VPD)
status and general trend over the day (solid
line).
12
Yepez et al. (2003)

• Vertically-distrib ?D and ?18O in semi-arid
  savanna woodland
• Upper/lower profiles analysed total and
  understory flux
• Post-monsoon transp 85 total, grass 50
  understorey ET
• Total ET 3.5mm/d 2.5 (70) tree trans 0.5
  (15) grass

13
Williams et al. (2004)

• Vert distrib ?D Morocco olive orchard following
  100mm irrig
• Keeling vs sap flow (v. hard to get
  representative data)
• Trans/soil evap by isotope method within 4/15
  sap flow
• Transpiration pre-irrig 100, post-irrig 70-85

14
Complex Canopies

• How can use of isotopes add value to
  understanding of ET from a complex
  canopy/ecosystem such as Tumbarumba?

15
Atmospheric Boundary Layer

• First need to understand turbulent mixing
  processes in the canopy, and interactions with
  atmospheric boundary layer

16
ABL Structure and Turbulence
(Holtslag and Duynkerke, 1998)
Day
(Wyngaard, 1990)
Night
17
Vegetation Canopies

• The essential differences between turbulence in
  the canopy air space and that in the boundary
  layer above result from the sources and sinks of
  momentum and scalars that are spread through the
  canopy (Kaimal and Finnigan, 1994)

• Canopy turbulence is dominated by the large
  eddies that form in the intense shear layer
  confined to the crown or upper part of the canopy

18
Wind in Vegetation Canopies

• Similar behaviour over large range of obs/model
  canopies
• Wind-shear max canopy top
• Attenuation below, foliage density determines
  rate
• Canopy turbulence strongly inhomogeneous in
  vertical
• All momentum absorbed in upper part of canopy
  (c.f. constant stress layer above)
• Large momentum gradient required to sustain
  steady air flow against aerodynamic drag of
  foliage

19
Turbulence in Vegetation Canopies

• Skewness
• Measure of turbulent intermittency
• Zero in surface layer
• Canopy SKu ve SKw-ve
• Turbulence is dominated by intermittent downward
  moving gusts (large eddies)

(Kaimal and Finnigan, 1994)

• Spectral peaks
• Canopy peak positions constant
• Large eddies extend through whole depth of
  foliage and into the air above

20
Turbulence in Vegetation Canopies

• TKE budget
• Shear prodn peaks near canopy top
• Wake prodn high in upper third
• Turbulent transport sink of TKE at canopy top,
  source in lower canopy
• Lower canopy TKE not locally produced imported
  from above by large eddies
• Dissipation much higher than free stream wake
  and waving terms convert dominant large scale
  motions to smaller eddies

(Kaimal and Finnigan, 1994)

• Canopy turbulence dominated by canopy-scale
  large eddies
• Cool dry gusts displacing warm moist canopy air
  at all levels
• Counter-grad fluxes non-local mixing turb
  transport distributed sources
• Surface layer flux-profile mixing relationships
  (K-theory) are inapplicable in vegetation
  canopies

21
Turbulence in Tumbarumba

• Quiescent at night
• Strong in daytime (900-1500) ABL convective
  motions

22
Temperature in Vegetation Canopies

• Night
• lower canopy unstable strat - enhanced mixing
• upper canopy stable strat (no turb - dew
  formation possible)
• Tumbarumba slightly stable (suppresses mixing)
• Daytime
• crown max (sun on foliage), with stable strat
  below. But ve (counter-grad) flux, so bimodal
• Intermittent mixing by large eddies quiescent
  periods
• Tumbarumba rapid increase of whole profile in
  morning unstable for remainder of day

23
Humidity in Vegetation Canopies

• Night
• Tumbarumba. Saturated (gt80 at 70m, colder
  below), with slow decrease of whole profile
  dew/fog
• Morning
• Tumbarumba. Rapid increase of whole profile
  dew/fog re-evap as temp incr transpiration
  kicks in
• Day
• Negative gradient progressive decrease of whole
  profile dry air intrusion
• Transpiration (secondary maximum in crown)
• Large values near ground surface moisture in
  leaf litter after rain

24
Precipitation 1-20 March 2005
25
Humidity in Vegetation Canopies
Surface moisture in leaf litter after rain
26
Isotopes in Vegetation Canopies
Isotope gradients all day
27
Isotopes in Vegetation Canopies
Transp. -40 o/oo
Atmos. -150 o/oo
Soil evap. -95 o/oo

• Night. ve grad condensation onto surface/plants
  (temp dep)
• Morning. Re-evap of (heavy) dew/fog transp
  soil evap
• Afternoon. -ve grad transp soil evap mixing
  from above

28
Vertical Keeling Analysis
Transp. -40 o/oo
Soil evap. -95 o/oo
Atmos. -150 o/oo
29
Tumbarumba Keeling Analysis

• Intercept from Keeling plots ?DET
• Guesses for ?D source values
• Soil evap -950
• Transpiration -40
• Total FT()
• n/a at night
• 20 morning (dodgy)
• 80 afternoon
• Understorey
• 60 at night (no!)
• 20 morning (dodgy)
• 50 afternoon

30
Tumbarumba Keeling Analysis

• r2 values only high in afternoon

31
Time-varying Keeling Analysis
Transp. -40 o/oo
Intercept -66.6 R20.762
Soil evap. -95 o/oo
Atmos. -150 o/oo
32
Conclusions

• Vertically varying SWI data can be used to add
  value to our understanding of moisture exchange
  in plant canopies, particularly the partitioning
  of evapotranspiration
• The combination of time-varying and
  vertically-varying mixing analyses
  (Keelingbetter?) of both ?D and ?18O promises to
  be a very powerful tool for analysing ET in
  complex ecosystems such as Tumbarumba
• But
• Need to understand the whole picture in terms
  of the airflow/turbulence regime within and above
  the canopy, so supporting meteorological data is
  essential.