data screening/ cleanup

[teixeirak]

check outlier / suspicious values:

• Heshan Hilly Land Interdisciplinary Experimental Station | Aggrading Eucalyptus urophylla Forest. Stand established around 2004 (NEP outlier)
• Haw River | Aggrading Temperate Evergreen Forest. Stand established around 1958.5 (NEP outlier)
• Arevalo_2011_luce research site in Alberta Canada
• Gainesville
• Turkey Point sites
• additional high Bond-Lamberty NEP values... --> these values are correct
• SRDB biomass values (see below)
• deadwood standing outlier (see below)
• deadwood down outlier (see below)
• fine root biomass outliers from SRDB (here)

check/ merge duplicates

• Caxiuana~ National Forest Reserve | Aggrading Tropical Evergreen Forest. Stand established around 1960. This has NEE record. (discussed here)

[ValentineHerr]

• additional high Bond-Lamberty NEP values..

looking into this now (citation Stape_2008_paca), looks like unit issues, not only with NEP.

[bpbond]

Let me know please! Thanks @ValentineHerr

[ValentineHerr]

For data pulled from Stape_2008_paca, coming from SRDB,

The values in ForC appear correct for:

• NEP (except for a decimal point, see values in main text, see quote below)
• ANPP_0 (ANPP values in Table 2)
• GPP (GPP values in Table 2)
• R_soil (F_s values in Table 2)

The conversion from kg C m-2 yr-1 to Mg C ha-1 yr-1 (multiplying by 10) did not happen for:

• TBCF (TBCA values in Table 2)
• ANPP_litterfall (F_a values in Table 2)

@bpbond are these 2 variables in kg C m-2 yr-1 in SRDB? if yes, you don't need to worry (and I need to fix ForC ), if not, they need to be fixed in SRDB (and I'll fix only these values in ForC).

In conclusion for this particular issue about NEP, unless the paper itself has wrong units, the high values of NEP in ForC are correct (except we ave 23.7 instead of 23.0) . They are these following values multiplied by 10 to get in units of Mg C ha-1 yr-1 :

Net ecosystem production increased with irrigation both in the wet year (2.3 kg C m2 year1 versus 2.7 kg C m2 year1) and
normal year (0.8 kg C m2 year1 versus 2.0 kg C m2 year1)

[bpbond]

SRDB units should be gC/m2/yr. So, you think they got entered as kgC instead?

[ValentineHerr]

yes. This the table in the paper. ANPP, GPP and F_s (R_soil in ForC) are these values x 10 but TBCA (TBCF in ForC) and F_A (ANPP_itterfall in ForC) are these exact values:

[ValentineHerr]

oh wait, @bpbond, I miss read you message... let me think more

[ValentineHerr]

@bpbond, in SRDB_data.csv, you entered 214, 246, 110 and 137 for values in the table above that are 2.14, 2.46, 1.10 and 1.37...
if you wanted gC/m2/yr I believe you should have entered 2140, 2460, 1100 and 1370.
These are records 1551-1554 in SRDB.

The same must have happened for ANPP_litterfall, but I have not looks at those records in SRDB

[bpbond]

OK thank you! Do you need me to do anything? I'm opening an issue for this over in SRDB.

[ValentineHerr]

no, thanks! I'll just edit in forC.

[ValentineHerr]

• Caxiuana~ National Forest Reserve | Aggrading Tropical Evergreen Forest. Stand established around 1960. This has NEE record.

@teixeirak, I don't see duplicates in this plot... nor do I see NEE values... did you fix that already? note that Caxiuana~ National Forest Reserve gets flagged to "revisit by hand" by the script creating PLOTS (see this issue)

[ValentineHerr]

The high R_soil values in this plot are form this paper and correct:

"The annual totals for Rs during 2008–2011
were 3.20, 3.89, 3.52, 4.14 kgCO2 m−2 years−1
, respectively"

--> values multiplied by 10 to get in MgC/ha/yr are in the 30's to 40's...

Maybe the values are high because, eventhough it is a young forest, it was a preserved forest before the clearing, so the soil were "old" ??

"The forest has been preserved since 2005,
when it was subjected to wood harvesting for charcoal
production. As a result, trees were mostly in a regeneration phase. In 2011, the trees were 6–7 years old with
average height and trunk diameter of 5.97 m and 6.59 cm, respectively. "

@teixeirak, don't we have a field in ForC to say that the measurement was double-checked?

[teixeirak]

@teixeirak, don't we have a field in ForC to say that the measurement was double-checked?

Yes. checked.ori.pub in MEASUREMENTS.

[teixeirak]

The high R_soil values in this plot are form this paper and correct:

"The annual totals for Rs during 2008–2011
were 3.20, 3.89, 3.52, 4.14 kgCO2 m−2 years−1
, respectively"

--> values multiplied by 10 to get in MgC/ha/yr are in the 30's to 40's...

Thanks for checking! There actually is an error here, though: original units were CO2, so we need to multiply by (12/44) to get C. I fixed in ForC.

@bpbond , there's another to fix in SRDB.

[teixeirak]

These two AGB records are definitely wrong:

[teixeirak]

For Proidencia values, see this issue. I deleted the whole study from ForC.

[teixeirak]

There are actually a number of AGB records from SRDB that can't be right (@bpbond ):

I don't have time to dig into these right now, but at least noting.

• 3 Deng 2012 Campoe 2012, Butler 2012 outliers were a missing decimal point. All values should be divided by 10 (@bpbond )
• Yashiro_2012_teod was wrong (not sure what happened), corrected.

[teixeirak]

Here's the biggest deadwood_standing outlier:

• could not locate the original study, flagged as suspicious

[bpbond]

Thanks!

[teixeirak]

Here are the biggest deadwood_down outliesr (but note bigger value if we get additional GROA sites classified)

• Tapajos is taken care of.
• Northern Chiloe Island is taken care of (but needs to be fixed in GROA, @CookPatton )
• Barangay Gaas is accurate.

[teixeirak]

These are 3 suspicious root biomass values from SRDB (huge outliers in fine root biomass age trends figure):

[teixeirak]

These are 3 suspicious root biomass values from SRDB (huge outliers in fine root biomass age trends figure):

I don't have access to this pub. Flagged as suspicious so they'll be excluded.

[teixeirak]

I've dealt with the current round of outliers. I'm going to close this and open another if more pop up.