Three-dimensional flows and NEE: Results from the Chequamegon Ecosystem-Atmosphere Study

(ChEAS)

Ken Davis, Weiguo Wang, Chuixiang Yi and others, The Pennsylvania State University

Paul Bolstad, Bruce Cook and Jon Martin, University of Minnesota

Peter Bakwin, NOAA/CMDL

Jud Isebrands and Ron Teclaw, USDA-FS

WISCONSIN

Coniferous

Mixed deciduous/coniferous

Wetland

Open water

Shrubland

General Agriculture

Willow Creek

WLEF

Lost Creek

Landcover key

North

Upland, wetland, andvery tall flux tower. Oldgrowth tower to the NE.

High-precision CO2

profile at each site.

Mini-mesonet, 15-20kmspacing between towers.

Progress• HNFs (humungous nighttime fluxes) identified

and isolated. Very large contribution to NEE of CO2. Regional vent of nighttime drainage flows?

• Advection terms computed from the ChEAS “mesonet” of CO2(x,z). Nighttime contributions of ~10%

• Multiple flux levels at WLEF yield upper limit of ~20 gC m-2 yr-1 uncertainty in annual NEE of CO2.

Humungous nighttime fluxes

• Peak fluxes (turbulent flux, not rate of change of storage) of ~ 80 mol m-2 s-1 detected under easterly winds and light stability. Strictly limited to easterly winds. Fluxes persist for several hours.

• Impact on cumulative NEE of CO2 is large. If screened, annual sum ~ - 400 gC m-2 yr-1. No screening, sum ~ - 130 gC m-2 yr-1.

Hypothesized flow

Computing net ecosystem-atmosphere exchange (NEE)

Ci

i

i

i Sx

CU

x

CU

t

C

''

turbulentstorage

advectionturbulentstorage

zs

zs

zs

FFNEE

FFFNEE

dzx

Cu

z

CW

x

CU

Cwdzt

CNEE

0

0

''

''

0

Mesonet advection calculations

• Use only hours with winds along the line defined by WLEF and Willow Creek.

• Assumes that CO2(x) is well-described by two points separated by ~15km.

• Vertical advection computed both as a residual, and directly via sonic mean vertical velocity and WLEF CO2(z).

• Hourly sonic resolution of 0.06 m s-1 determined from two dual-sonic deployments.

Mesonet advection calculations

Integrated effect on NEE is a 10% underestimate of nighttimeflux. Order 50 gC m-2 yr-1 if extrapolated over a year

June-August 1997 diurnal mean cumulative NEE at WLEF vs. level

2.211.151.011.0019-4

-0.581.161.011.00415-18

-2.011.060.940.9011-14

-1.871.161.290.985-10

0.960.98All Day w/ -9

-2.251.081.180.89All Day

Preferred gC m-2 d-1

396m122m30mHours (LST) (Fraction of preferred NEE)

1997 Cumulative NEE, GEP and RE vs. assumptions and methods

1733-1758-25 +/- 17Low U* screened, median fill

1634-1681-48 +/- 20Low U* retained

1924-190916 +/- 19Low U* screened,

T-PAR fill

REGEPNEEMethod(gC m-2 yr-1 = tC ha-1 yr-1 * 100)

WLEF summary• WLEF region 1997 annual NEE is about 0.• Identified systematic uncertainties

– Different levels: footprint/advection – order 20 gC m-2 yr-1

– U* screen – order 50 gC m-2 yr-1

– Wind direction – didn’t appear to be large– But surface energy balance isn’t obtained.

• Random errors (weather + sampling)– Order 20 gC m-2 yr-1.

• GEP and RE values are very significant