Regional estimates of NEE and comparisons with tower measurements

1
Regional estimates of NEE and comparisons with
tower measurements --Updates and
summary Department of Meteorology, Pennsylvania
State University June 1, 2005
2
Outline 1. Motivations 2. Methods - Flux
aggregation - ABL budget method 3. Results
and Comparison
3

• Motivations
• Implications from the rejection of the initial
  hypothesis (WLEFWCLC weighted by area)
• Estimate regional NEE over this heterogeneous
  ecosystem

4
Site
SourceWISLAND
5
Data used - WLEF fluxes and co2 mixing ratio
profile - Willow Creek fluxes - Lost Creek
Fluxes - others
6

• Flux aggregation
• Watershed function level
• Wetland and Upland
• Stand type level, 6 types
• 1 Mixed coniferous, deciduous upland
  forests
• 2 Aspen
• 3 other deciduous forest
• 4 Lowland wetland
• 5 Forested wetland
• 6 Others

Upland wetland
7

• Question How do we know fluxes for more
  ecosystem types when sufficient measurements are
  not available?
• A follow-up question Are fluxes (inferred
  and measured) representative of the region?
• Method Decompose the fluxes measured over the
  mixed ecosystems to infer fluxes for more types.

8
Decomposing Method

• Based on footprint theory, fluxes measured at the
  WLEF tower can be interpreted as a weighted
  average of NEE of all types in the footprint
  area,

Weight for ecosystem type i footprint
NEE for ecosystem type i n-total types NEEf(PAR,
p) f(T, p)
Measured flux at height zm and time t
Note that, In the day, effects of entrainment at
CBL top is ignored. For low levels, small.
9
Growing-season average daily fluxes
ER GEP NEE
Implications (1) Different responses of
ecosystem types within forested uplands, and
within wetlands (2)Different responses of
similar ecosystem types at different sites,
implying that more detailed classification
schemes are needed.

Forested uplands Wetlands
10
Daily Fluxes in growing season
ER
Difference in aggregated NEE can be as large as
400 gC/m2/season Suggest that it is important to
distinguish wetland upland, within wetland
category Within upland forests category

GEP
NEE
WCLC Extended WLEF
WLEF/WC/LC
WLEF tower alone
11
ABL budget method

• Benefit from the long-term vertical profile of
  CO2 measured at the tall tower
• - direct measurements in the mixed layer
• - estimate CO2 jump
• - top level gt NBL top
• Estimate regional fluxes on long time scales

12
RegNEE 3 levels at WLEF tower

NEE at 396m is closest to Reg estimates
13
Monthly average ABL NEE WC, LC, WLEF

-Bounded by existing tower measurements at a
given PAR or at a given air temperature. -Differen
ce in max and min NEE at towers is so big, which
is close to magnitude of regional NEE. Tower
constraints are insufficient.
14
Nighttime NEE, WLEF is max in the growing
season, WC is max in the dormant season

15
Comparison with other estimates
--In the dormant season, WLEF measurement close
to NEE in the region studied (103km2), even in a
larger region (104-105km2) --This is not true
during the growing season. Significant Difference
in NEE among ecosystems
16
Comparison of annual cumulative NEE in 2003
NEE types Annual Cumulative gC/m2/yr Suggested net Source/sink
regNEE -175 ?60 sink
WLEF 80?16 source
WC -502 Big sink
LC 10 Small source
17
Future work

• Quantify footprints under inhomogeneous
  conditions (surface, turbulence) better
  interpret the measurements
• Quantify the uncertainty of the flux
  measurements. (not only the magnitude of order)
• More measurements are needed to either confirm
  the implications, or estimate the regional flux
  with more confidence.
• Vertical profile measurements at more than one
  location to reduce uncertainty due to
  assumptions.
• Others