1 LEVEL 4 – VARIABLE DESCRIPTION
5 Level 4 data are obtained from the level 3 products, data are ustar filtered, gap-filled using different methods and partitioned. Datasets are also aggregated from daily to monthly. Flags with information regarding quality of the original and gapfilled data are added.
7 Half hourly dataset variables description:
10 - Day: day of the month
11 - Hour: from 0 to 23.5, indicates the end of the half hour of measurement
12 - DoY: decimal day of the year
13 - Rg_f: global radiation filled [W m-2]
14 - Rg_fqc: global radiation quality flags: 0 = original, 1 = category A (most reliable), 2 = category B (medium), 3 = category C (least reliable). (Refer to Reichstein et al. 2005 Global Change Biology for more information)
15 - Ta_f: air temperature filled [°C]
16 - Ta_fqc: air temperature quality flags: 0 = original, 1 = category A (most reliable), 2 = category B (medium), 3 = category C (least reliable). (Refer to Reichstein et al. 2005 Global Change Biology for more information)
17 - VPD_f: vapour pressure deficit [kPa]
18 - VPD_fqc: vapour pressure deficit quality flags: 0 = original, 1 = category A (most reliable), 2 = category B (medium), 3 = category C (least reliable). (Refer to Reichstein et al. 2005 Global Change Biology for more information)
19 - Ts_f: soil temperature filled [°C]
20 - Ts_fqc: soil temperature quality flags: 0 = original, 1 = category A (most reliable), 2 = category B (medium), 3 = category C (least reliable). (Refer to Reichstein et al. 2005 Global Change Biology for more information)
21 - Precip: precipitation [mm]
22 - SWC: soil water content [%vol]
23 - H_f: sensible heat flux filled [W m-2]
24 - H_fqc: sensible heat flux quality flags: 0 = original, 1 = category A (most reliable), 2 = category B (medium), 3 = category C (least reliable). (Refer to Reichstein et al. 2005 Global Change Biology for more information)
25 - LE_f: latent heat flux filled [W m-2]
26 - LE_fqc: latent heat flux quality flags: 0 = original, 1 = category A (most reliable), 2 = category B (medium), 3 = category C (least reliable). (Refer to Reichstein et al. 2005 Global Change Biology for more information)
27 - qf_NEE_st: fluxes quality flags as defined in the Level3 product
28 - qf_NEE_or: fluxes quality flags as defined in the Level3 product
29 - Reco_st: Estimated ecosystem respiration according to the short-term temperature response of night-time fluxes based on NEE_st (Refer to Reichstein et al. 2005 Global Change Biology for more information) [umolCO2 m-2 s-1]
30 - Reco_or: Estimated ecosystem respiration according to the short-term temperature response of night-time fluxes based on NEE_or (Refer to Reichstein et al. 2005 Global Change Biology for more information) [umolCO2 m-2 s-1]
31 - NEE_st_fMDS: NEE_st filled using the Marginal Distribution Sampling method (Refer to Reichstein et al. 2005 Global Change Biology for more information) [umolCO2 m-2 s-1]
32 - NEE_st_fMDSqc: NEE_st_fMDS quality flags: 0 = original, 1 = category A (most reliable), 2 = category B (medium), 3 = category C (least reliable). (Refer to Reichstein et al. 2005 Global Change Biology for more information)
33 - GPP_st_MDS: Gross Primary Production calculated as GPP_st_MDS = Reco_st - NEE_st_MDS [umolCO2 m-2 s-1]
34 - NEE_or_fMDS: NEE_st filled using the Marginal Distribution Sampling method (Refer to Reichstein et al. 2005 Global Change Biology for more information) [umolCO2 m-2 s-1]
35 - NEE_or_fMDSqc: NEE_or_fMDS quality flags: 0 = original, 1 = category A (most reliable), 2 = category B (medium), 3 = category C (least reliable). (Refer to Reichstein et al. 2005 Global Change Biology for more information)
36 - GPP_or_MDS: Gross Primary Production calculated as GPP_or_MDS = Reco_or - NEE_or_MDS [umolCO2 m-2 s-1]
37 - NEE_st_fANN: NEE_st filled using the Artificial Neural Network method (Refer to Papale et al. 2003 Global Change Biology for more information and to the Other Information section in this document) [umolCO2 m-2 s-1]
38 - NEE_st_fANNqc: NEE_st_fANN quality flags: 0 = original, 1 = filled using original meteorological inputs or filled with qc=1, 2 = filled using filled meteorological inputs with qc=2 or 3, 3 = not filled using ANN due to one or more input missed but filled with the MDS method
39 - GPP_st_ANN: Gross Primary Production calculated as GPP_st_ ANN = Reco_st - NEE_st_ ANN [umolCO2 m-2 s-1]
40 - NEE_or_f ANN: NEE_or filled using the Artificial Neural Network method (Refer to Papale et al. 2003 Global Change Biology for more information and to the Other Information section in this document) [umolCO2 m-2 s-1]
41 - NEE_or_f ANNqc: : NEE_or_fANN quality flags: 0 = original, 1 = filled using original meteorological inputs or filled with qc=1, 2 = filled using filled meteorological inputs with qc=2 or 3, 3 = not filled using ANN due to one or more input missed but filled with the MDS method
42 - GPP_or_ ANN: Gross Primary Production calculated as GPP_or_ ANN = Reco_or - NEE_or_ ANN [umolCO2 m-2 s-1]
46 Daily to Monthly aggregated dataset variables description:
48 - Month (in Daily and Monthly files): from 1 to 12
49 - Day (in Daily file): day of the month
50 - DoY (in Daily file): day of the year
51 - Period (in Weekly file): 8-days period from 1 to 46
52 - n_days (in Weekly and Monthly files): number of days used to calculate the mean daily value associated with the month or 8-days period
53 - Rg_f: global radiation filled [MJ m-2 day-1]
54 - Rg_sqc: global radiation summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
55 - Ta_f: air temperature filled [°C]
56 - Ta_sqc: air temperature summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
57 - VPD_f: vapour pressure deficit [kPa]
58 - VPD_sqc: vapour pressure deficit summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
59 - Ts_f: soil temperature filled [°C]
60 - Ts_sqc: soil temperature summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
61 - Precip: precipitation [mm day-1]
62 - SWC: soil water content [%vol]
63 - H_f: sensible heat flux filled [W m-2 day-1]
64 - H_sqc: sensible heat flux summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
65 - LE_f: latent heat flux filled [W m-2 day-1]
66 - LE_sqc: latent heat flux summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
67 - Reco_st: ecosystem respiration based on NEE_st [gC m-2 day-1]
68 - Reco_or: ecosystem respiration based on NEE_or [gC m-2 day-1]
69 - NEE_st_fMDS: NEE_st filled using the Marginal Distribution Sampling method [gC m-2 day-1]
70 - NEE_st_fMDSsqc: NEE_st summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
71 - GPP_st_MDS: gross primary production based on NEE_st filled with the Marginal Distribution Sampling method [gC m-2 day-1]
72 - NEE_or_fMDS: NEE_or filled using the Marginal Distribution Sampling method [gC m-2 day-1]
73 - NEE_or_fMDSsqc: NEE_or summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
74 - GPP_or_MDS: gross primary production based on NEE_st filled with the Marginal Distribution Sampling method [gC m-2 day-1]
75 - NEE_st_fANN: NEE_st filled using the Artificial Neural Network method [gC m-2 day-1]
76 - NEE_st_fANNsqc: NEE_st summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
77 - GPP_st_ANN: gross primary production based on NEE_st filled with the Artificial Neural Network method [gC m-2 day-1]
78 - NEE_or_fANN: NEE_or filled using the Artificial Neural Network method [gC m-2 day-1]
79 - NEE_or_fANNsqc: NEE_or summary quality flag, indicates the percentage of half hourly data with qc = 0 or 1 used in the aggregation
80 - GPP_or_ANN: gross primary production based on NEE_st filled with the Artificial Neural Network method [gC m-2 day-1]
83 Please note that the qc flags associated with NEE filled using ANN are not exhaustive because the quality depends also by other aspects like the availability of data in the same period of the year and of the day that can be used for the training process.