Dear researchers,
I am working on CO2 flux data measured at two different heights (~3 m and ~40 m) at a tower in the middle of a large homogeneous area (cropland, 6km x 6km). Anthropogenic emissions are ignorable. I am curious why CO2 fluxes (or NEE) showed a significant difference between two different measurement levels in daytime. The lower the larger absorption by ~ 2 times around noon LST (See the attached Figure). I think the two fluxes should be equal or similar, because their sources are almost equal, except the area covered by footprints, but the figure attached showed a huge offset.
A few other ideas come up:
1) I set a criteria of u* > 0.2 m/s. Should I adjust or found the proper u* at each measurement level? But the estimation of U* threshold distribution did not likely give any clear answer.
2) Is the higher one not sit on the constant-flux layer? Then, how can people interpret the very tall tower (> 200 m) data over a forest area? What is the main difference of fluxes measured at the canopy-level and the much-above level?
3) Should I add storage term to each flux to achieve NEE? I did, but the still large amount of offset.
In sum, how does one interpret the different fluxes at different measurement level at the same homogeneous site? Anyone can help me out?
Regards,
I dont think they can be the same. many factors could be responsible for the difference in results e.g wind speed and direction which can be influenced by the density of buildings in ther area etc
Use several times the classic chamber at several spots within a tower footprint. It is not a perfect technique but direct one.
As pointed out earlier, variability in mixing of gases at different heights from ground may affect the NEE. For example at a region where the average canopy height is 3 m, the 3m tower would capture more instantaneous shifts in NEE as opposed to the 40 m. If you go below the canopy height and closer to ground you may get more CO2 efflux than influx and hence a more positive NEE as carbon uptake may decrease and emissions from soil increase. Finally, it all depends on how well the gases mix (both vertically and horizontally) at different layers.
Hi Changhyoun,
To explain your observation I must refer to the basics of aerodynamics and eddy covariance measurement modelling. Specifically, the term flux footprint describes an upwind area determining an instrument response to vertical turbulent fluxes of heat, mass (water, gas (CO2)) and momentum generated in a specific area, the footprint.
Three main factors affect the size and shape of the flux footprint:
- Measurement height! (above for example homogeneous vegetation);
- Surface roughness;
- Atmospheric (thermal) stability.
See more at the following URL:
https://en.wikipedia.org/wiki/Flux_footprint#Main_factors_affecting_flux_footprint
Cheers,
Frank
Hi Changhyoun!
Very large difference there! What my colleagues above highlight should normally NOT create such a large discrepancy. Flux footprint is likely most relevant, but also atmospheric layering. Your nighttime data suggest that some storage is most likely happening (little to no storage during daytime).
CO2 flux should decrease with height above the constant flux layer, but not this much by a factor of 2 at 40 m vs. 3 m, I think. Your 3-m NEE is much larger (not smaller), so it is not likely local respiration, but local Pn vs. larger scale Pn. Your flux footprints are much different at those heights!
Some possibilities:
- if roughness length is very low (what is the crop?), the z=40 m flux footprint may extend substantially outside the cropland area
- the concentration footprint for z=40 m may extend substantially outside the cropland area and you may see a horizontal inhomogeneity effect from a large, distant CO2 source, such as a power plant. Obviously you can check for that using wind direction discrimination ...
- your field(s) may not be as homogeneous as you think ... strong gradients may exist in Pn or respiration depending on soil moisture ... how do your latent heat fluxes compare? are there any canopy gaps?
- if you have access to a FLIR camera, check what the crop looks like in the IR; that can give clues on heat stress (you may sit in a local spot of optimal Pn ...)
Good luck figuring this out,
Gunnar
I really appreciate all of you for responding to my discussion and helping me lead to further understanding.
The wind direction analysis revealed some sources in E directions within the 40-m footprints. After removing the source effects, I see still the decreasing pattern with height but not that much different as in the previous figure. Maybe the higher level is affected by anthropogenic sources. Thanks to Gunnar's comments.
Here is my conclusion: NEEs can be measured independent of measurement heights as long as their footprints cover homogeneous areas and their measurement points are within the constant-flux layer.
Additional comments are always welcome.
Hi Changhyoun,
Others already asked about your canopy and its roughness. If you have measurements at 3 m, I can guess it could be some kind of field crop or grass. Would be nice to know what you have below your tower. More information is required to understand the footprints seen by the two systems.
You mentioned your worry about sitting in the constant flux layer with the upper level, but - maybe - you are in the roughness sublayer with the lower... There, enhanced turbulence may affect your eddy flux in a quite unpredictable way.
I would like to point also at field levelling and the coordinate rotation, that is routinely applied to wind vector. Usually, this is "automatically" done by processing packages, and one does not care much. Did you rotate data to get fluxes?
Interesting question, by the way!
Best, Andrea
Thank you, Andrea. I forgot to provide the site info.
The tower is in the middle of a rice paddy field. I used the EddyPro software to handle the 10-Hz raw data.
I thought the 3-m height is OK for the flux measurements. But as you mentioned, if the lower level would be in the roughness sublayer not in the constant flux layer, what am I supposed to do further? Should I discard the lower measurement data and use only the upper levels for NEE estimations?
Hi Changhyoun Park,
If you have a rice field below your tower, I would say that 3m is more than appropriate, and your measurements should not suffer from being too close to the roughness sublayer.
I have not any experience with rice (and I do not know how developed and leafy it was...), but I would expect -10 umol/m2/s as a very convincing midday flux for a fully functional and healthy crop canopy. Also very convincing (good mixing) seem to be the nightime trends for 3m. Less convincing - actually - are the nighttime trends for the higher level. See, for instance, the penultimate point, that shows some net absorption still during nightime... Very obviously, the 30m instruments are quite decoupled from the surface, especially under stable conditions.
All in all, I would trust more the 3m measurements, and I would try to understand why the higher is "off". Footprint is the major issue, and on such a smooth (and flat, I suppose...) surface, with low sensible heat fluxes (flooded rice...), it can extend well away upwind your field. EddyPro can tell you how far. Have a check, and see if you still have your canopy there.
Best, Andrea
Very interesting discussion, adding further to Andrea's and Gunnar's comments, I would suggest to check the homogeneity of the 40 m footprint area with Sentinel-2 ndvi and ndwi data. As previously mentioned, it is very probable that the field is not as homogeneous as it may appear from the ground.
Some information about this question may be found at link Preprint Formation of CO and CO2 molecules under the tree roof and th...