Management Data#
Info
Management information was converted to a time series format with the aim to be used e.g. in machine-learning models.
Notebook
Download management info as Excel file
CH-CHA-Management_2005-2024_20250107 (zipped XLSX file, 8MB)
This file contains detailed management data, with info between 2001 and 2024. At the time of writing, some info (e.g., C content of fertilizer) was not yet available for the most recent year.
Unique events#
Each management event was assigned a variable name and a list of unique events was assembled. In total, we identified 34 unique managmement events.
Unique management events |
|---|
1: MGMT_FERT_ORG_MAN |
2: MGMT_PPGM_HERBREMOVAL_RUMEX |
3: MGMT_PPGM_HARROWING_1CM |
4: MGMT_USE_GRAZING_CATTLE |
5: MGMT_PPGM_ROLLING |
6: MGMT_PPGM_HARROWING_7CM |
7: MGMT_SOWING_OVERSOWING_OH440RENO |
8: MGMT_USE_MOWING_GRASSSILAGE |
9: MGMT_FERT_MIN_CAN |
10: MGMT_FERT_ORG_SLU |
11: MGMT_USE_GRAZING_SHEEP |
12: MGMT_SOWING_OVERSOWING_OH440EXTRA |
13: MGMT_USE_MOWING_AFTERGRASS |
14: MGMT_PPGM_HERBICIDE_ASULOX |
15: MGMT_SOILCULTIVATION_ROLLING |
16: MGMT_USE_MOWING_HAY |
17: MGMT_SOWING_OVERSOWING_OH440RENOTURBO |
18: MGMT_SOWING_OVERSOWING_OH240RENO |
19: MGMT_PPGM_HERBICIDE_ALLYTABS |
20: MGMT_SOILCULTIVATION_PLOUGHING |
21: MGMT_SOILCULTIVATION_HARROWING_7CM |
22: MGMT_SOWING_RESOWING_OH440EXTRA |
23: MGMT_PPGM_PESTICIDE_MOLLUSCICIDE |
24: MGMT_PPGM_HERBICIDE_HARMONYTABS |
25: MGMT_PPGM_HERBICIDE_ASULAM |
26: MGMT_SOWING_OVERSOWING_HEBE-FIONA-TEDI |
27: MGMT_SOILCULTIVATION_HARROWING_1CM |
28: MGMT_USE_MOWING |
29: MGMT_SOWING_OVERSOWING_HEBE-FIONA-MERULA |
30: MGMT_PPGM_HERBICIDE_ROUNDUP |
31: MGMT_SOILCULTIVATION_TILLAGE_3-4CM |
32: MGMT_SOWING_RESOWING_UFA-420 |
33: MGMT_PPGM_HERBICIDE |
34: MGMT_SOILCULTIVATION_HARROWING |
Simplified variable names#
Unique events were simplified and grouped together, where appropriate. In addition, the parcel info was added to the variable names. We identified 14 events that were relevant for either parcel A or parcel B.
Management per parcel |
counts |
|---|---|
MGMT_MOWING_PARCEL-B |
118 |
MGMT_FERT_ORG_PARCEL-A |
112 |
MGMT_MOWING_PARCEL-A |
110 |
MGMT_FERT_ORG_PARCEL-B |
85 |
MGMT_GRAZING_PARCEL-B |
36 |
MGMT_GRAZING_PARCEL-A |
34 |
MGMT_SOWING_PARCEL-B |
21 |
MGMT_SOILCULTIVATION_PARCEL-B |
17 |
MGMT_PESTICIDE_HERBICIDE_PARCEL-B |
12 |
MGMT_PESTICIDE_HERBICIDE_PARCEL-A |
11 |
MGMT_SOILCULTIVATION_PARCEL-A |
11 |
MGMT_SOWING_PARCEL-A |
6 |
MGMT_FERT_MIN_PARCEL-A |
5 |
MGMT_FERT_MIN_PARCEL-B |
2 |
Converting management info to time series format#
Daily time scale#
In order to convert start and end information for each event to time series format, an empty dataframe with daily time resolution was created. Since the management info goes back to 2001, the timestamp of the dataframe was extended. As such, the first timestamp was 2001-01-01, the last timestamp 2023-12-31.
The info from the Excel file was then inserted into the empty dataframe. When management took place, values in the respective data column for the respective day(s) were set to 1, otherwise to 0. All values of the respective days of management were set to 1 because the exact starting and end times were not available for all events.
Example
The event MGMT_GRAZING_PARCEL-B had the start date 2008-11-01 and the end date 2008-11-04. In the dataframe, 1 was inserted in the column MGMT_GRAZING_PARCEL-B between the (daily) timestamps 2008-11-01 and 2008-11-04 (inclusive start and end dates).
For all management events, TIMESINCE variables were calculated, describing the temporal distance of each day to the previous management event.
XXX TODO: add figure
Example
The event MGMT_GRAZING_PARCEL-B had the start date 2008-11-01 and the end date 2008-11-04. The variable TIMESINCE_MGMT_GRAZING_PARCEL-B was calculated and has value 0 between 2008-11-01 and 2008-11-04, value 1 on 2008-11-05, value 2 on 2008-11-06, etc…
Half-hourly time scale#
Flux and meteo data are available with a half-hourly (hh) timestamp. Therefore, the daily management dataframe was converted to hh time resolution. The hh timestamp shows the middle of the averaging interval, TIMESTAMP_MIDDLE.
Example
At the daily scale, the event MGMT_GRAZING_PARCEL-B had the start date 2008-11-01 and the end date 2008-11-04. Converted to the half-hourly timescale, all values between 2008-11-01 00:15:00 and 2008-11-04 23:45:00 were set to 1.
Wind direction was then added to the hh management data and used to create new _FOOTPRINT (suffix) variables. These variables contain info about the parcel from where the wind was arriving at the sensors. This means, depending on wind direction, the _FOOTPRINT variables can contain info from _PARCEL-A or _PARCEL-B variables.
Parcel division runs diagonally from 250° to 70°, with the measurement station at the center. In earlier years (2005-2013), the division was less clear, with the division line shifted further to the south, but an analysis in Feigenwinter et al. (2023) showed that most of the contribution to the flux footprint (>97%) came from PARCEL-B, north of the station.