Table 2.
Summary of predictor variables used for machine learning modeling
No. | Product and its description | Predictor’s | Data | Data |
---|---|---|---|---|
type | source | resolution | ||
1 | Altitude as derived from digital elevation model (DEM) | Spatial | SRTM-3 | ca. 1 km |
2–3 | Geographical coordinates (in the projected coordinate system) | Spatial | – | |
4 | Distance to the Baltic Sea coast | Spatial | – | |
5–16 | Monthly mean air temperatures (Jan–Dec) | Meteo | E-OBS | ca. 27 km |
17 | Monthly mean air temperatures of previous’ year December | Meteo | E-OBS | ca. 27 km |
18–21 | Seasonal mean air temperatures of previous’ year | Meteo | E-OBS | ca. 27 km |
22–23 | Mean air temperatures of winter and spring seasons | Meteo | E-OBS | ca. 27 km |
24–35 | Total monthly precipitations (Jan–Dec) | Meteo | E-OBS | ca. 27 km |
36 | Total monthly precipitations of previous’ year December | Meteo | E-OBS | ca. 27 km |
37–45 | Cumulative growing degree days (GDD) | Meteo | E-OBS | ca. 27 km |
from 0 to 8 ∘C with an interval of 1 ∘C | ||||
46 | Cumulative growing precipitation days (GPD) | Meteo | E-OBS | ca. 27 km |
47 | Presence of snow cover (0–1) | MODIS | IMS | 4 km |
48–49 | Consecutive number of days with and without snow cover | MODIS | IMS | 4 km |
50 | Number of days with snow cover in a month | MODIS | IMS | 4 km |
51 | Day of year with the last snow cover | MODIS | IMS | 4 km |
NDVI, EVI, LAI, and fPAR means aggregated in AtPol 10 × 10-km grid: | MODIS | MYD13Q1 / | 0.25–1 km | |
MOD13Q1 | ||||
52–55 | - For all available values | MODIS | – || – | 0.25–1 km |
56–59 | - Based only on the highest pixel reliability (i.e., flagged as “0”) | MODIS | – || – | 0.25–1 km |
60–63 | - Based only on the highest and average pixel reliability (“0–1”) | MODIS | – || – | 0.25–1 km |
64–67 | - Based on all available except the lowest pixel reliability (“0–2”) | MODIS | – || – | 0.25–1 km |
68–79 | - 1-week rolling mean group by pixel reliability (“0,” “0–1,” “0–2”) | MODIS | – || – | 0.25–1 km |
- The rate of change grouped by pixel reliability (“0,” “0–1,” “0–2”) between: | ||||
80–91 | – Monthly and 10-day average | MODIS | – || – | 0.25–1 km |
92–103 | – 10-day average and one-week rolling mean | MODIS | – || – | 0.25–1 km |
104–115 | - Normalized for particular pixel’s location and | MODIS | – || – | 0.25–1 km |
grouped by pixel reliability (“0,” “0–1,” “0–2”) |