ASTM6 Early Career Researcher Poster Session 1 Chat Record
- https://above.nasa.gov/meeting_2020/ASTM6_ECR_Posters.pdf
- Eric - What do you think the future trajectory of the Arctic carbon budget will be - increasing source, increasing sink, or close to net neutral?
- What is the time duration and domain of the ED model runs?
- ED spun up from 1800-1979, then used NARR meteorology from 1979-2018. Then used RCP85 for 2019-2100. We see that out to 2100 tundra vegetation increases more than tundra soil carbon loss, however, in the boreal soil carbon loss is dominant. Over the ABoVE domain, the net change is ~1 kg/m2 carbon loss to the atmosphere
- Erik - do you only have results over ABoVE domain? or your runs are for the pan-Arctic?
- Which model is it?
- So that 1 kg/m2 is the mean integrated C loss for the 80 years from 2019-2100, correct? The ABoVE domain is approx 4E6 km2 --> 4E12 g net C loss by 2100. I
- So far we have just run the ABoVE domain. We could run the pan Arctic, if there is interest however.
- Erik, do you have plan to run ED for other scenarios (RCP 2.6, 4.5 etc)?
- Sorry - 4E12 kgC net loss = 4E15 g (Pg)
- That is correct Chip. We have an RCP 4.5 simulation as well. I did not show the results on our poster. Interestingly, since both GPP and RH increase with warming, the net change was not dramatically different between the simulations.
- Erica - what is the primary causes on the poor relationship between GPP and SIF?
- We think a lot of it has to do with canopy heterogeneity, and also perhaps not having enough flights over several days to average out the noise a little bit better.
- Have you done a spatial analysis of the variability in SIF? Specifically, what is the length scale variabiliity of SIF?
- Could the worse agreement at low GPP be related to uncertainties in the tower partitioning?
- Colm, not yet but we definitely will be doing that in the future. Length meaning wavelength?
- Roisin, yes that is also a possibility for sure, we need to dig further into the partioning methods for the different sites
- Hi Eric I was actually thinking about about the auto correlation lenght of SIF. You meantioned high variability and when comparing to flux tower data this would be an important parameter to quantify
- That's a good point, yes I will definitely look into that more. Thanks!
- I'm not sure if the CFIS data have been analyzed in the ABoVE domain yet for autocorrelation length, but it depends a bit on flight speed and altitude (ground resolution). It can easily resolve changes over indivdual agricultiural plots at the ha (100 m) scale
- those are good points Colm and Chip, thanks!
- What is relative SIF?
- relative SIF is SIF/reflected radiance in the retrieval window (far-red)....a closer approximation of ~SIFyield
- where is tower GPP for Delta Junction?
- Still being processed
- I'm working with the NEON folks to get the tower data for 2017 and then will add it and Toolik Lake into the analysis for mine.
- For tundra, GOME2 SIF was earlier than EVI but matched the NEE. For Boreal it was the other way around (EVI was earlier than SIF)...
- (Limited site level testing)
- Roisin - we can look at all these VIs and SIF from our tower data so that might help : )
- @Roisin, didn't Walther et al. 2016 suggest that SIF tracked startup better than EVI?
- Yes
- in boreal forest?
- Roisin you should take a look at Zoe Pierrats poster
- See DOI: 10.1111/gcb.14283 for Parazoo et al. Spring photosynthetic onset and net CO2 uptake in Alaska triggered by landscape thawing
- From Troy's comment, Bowling et al. 2018 -- which looks at spring awakening at Niwot Ridge
- also Esminger et al. 2004 GCB