Nexus Observatory Sign in

StoriesPlace: Farmland

Field to forecast

A farm observes water: in the soil, in the crop and in the air above it. Crop water use is computed from weather with a method agronomists have shared for decades1; what changes is how densely and how cheaply the inputs can be measured.

Systems
LPWAN soil sensing, Weather station, Machine data
Toward 2030
Field-scale Earth observation
Figures
3, field scene, simulator, layer stack

Go to the references

01Field scene

Probes, a gateway and a weather station.

Soil-moisture probes at root depth across the plots report a few bytes every quarter of an hour to one gateway on a mast. A weather station supplies radiation, temperature, humidity and wind, the inputs of reference evapotranspiration1.

Machinery in the field speaks its own bus2; the edge joins both so that an irrigation decision can cite the readings it was based on.

Figure as text: Plots, soil probes, a gateway mast, a weather station and a microgrid

Six ploughed plots in a wireframe landscape; four soil probes link by dashed red lines to a gateway mast; a weather station and a solar microgrid stand at the edges.

  • LPWAN gateway
  • Soil probe
  • Weather station
  • Microgrid

02Simulator

Few bytes, long range, little airtime.

A soil probe is limited less by range than by airtime: in licence-exempt bands, a device may only transmit for a small share of each hour3. Higher spreading factors reach further but stay on air longer, as the modem’s time-on-air formula shows4.

Figure as text: Time on air and uplinks per hour

A bar chart of time on air for spreading factors 7 to 12 with the chosen factor in red.

Time on air and uplinks per hour, at the starting values
Spreading factor9 SF
Payload18 bytes
Duty-cycle limit1 %
Time on air (125 kHz, CR 4/5)185.3 ms
Uplinks per hour within the limit194
Shortest interval between uplinks19 s

03Layer stack

Every layer of the field, registered.

Agronomy happens where layers meet: the terrain that routes water, the imagery that shows vigour, the soil classes, the probes, and the model that turns them into irrigation zones. They are useful only if they share one grid and one time base.

Figure as text: The layers of one field, from terrain to irrigation zones

An exploded stack of seven square plates: terrain contours, imagery strips, soil classes, drainage lines, probes with three live ones in red, an analysis grid and a model plate whose red isoline marks the zone to irrigate.

  1. Terrain: Contours from an elevation model.
  2. Imagery: Optical scenes resampled to the field grid.
  3. Soil classes: Mapped texture classes.
  4. Drainage: Flow lines derived from the terrain.
  5. Probes: Soil-moisture probes; red ones are reporting now.
  6. Analysis grid: Hexagonal cells for aggregation.
  7. Irrigation zones: Modelled soil-water deficit; the red line is the irrigation threshold.

Figure 1

Field scene

Gateway to probe horizon
21.2 km
Farthest probe
319 m

Figure 1. Plots, soil probes, a gateway mast, a weather station and a microgrid. Six ploughed plots in a wireframe landscape; four soil probes link by dashed red lines to a gateway mast; a weather station and a solar microgrid stand at the edges.

References

  1. Crop evapotranspiration: guidelines for computing crop water requirements (FAO Irrigation and Drainage Paper 56). Allen, R. G., Pereira, L. S., Raes, D. and Smith, M., Food and Agriculture Organization of the United Nations, 1998.
  2. ISO 11783 (ISOBUS): Tractors and machinery for agriculture and forestry — Serial control and communications data network. International Organization for Standardization, 2017.
  3. ETSI EN 300 220-2: Short range devices operating in 25 MHz to 1 000 MHz (duty-cycle limits). European Telecommunications Standards Institute, 2018.
  4. SX1276/77/78/79 datasheet, section 4.1.1.7: LoRa time on air. Semtech, 2020.

Related stories

System: LPWAN and mesh Kilometres on a coin cellMost environmental sensors send a few bytes a few times an hour. Low-power wide-area networks trade bandwidth for range and battery life; mesh networks…Horizon 2030: AI weather and geospatial foundation models Models that learned the atmosphereIn three years, machine-learned weather models went from research papers to operations. They learned from decades of reanalysis, which means they inherit…Place: Rural village A village at the end of the lineWhere the fibre ends, a single mast, a satellite terminal and a handful of sensors decide whether a village is on the map of observation. Canada’s target is…

Where this applies

Sensors & instrumentsConnectivityData & licensingEarth observation & GeoAIAgricultureWaterClimateField tour: farm