SOIL MOISTURE
31%Input value
Target 34% (demo)
FIELD STATUS
Good morning, YLN FM ☀
Sample, uploaded and historical inputs for prototype decision support.
Field at a glance
Input observations and calculated prototype indicators · no live sensors
WATER STRESS
LowCalculated
Unvalidated soil threshold rule
WATER BALANCE
−4.8mmSample sparkline
NEXT IRRIGATION
Monitor
No irrigation scheduled
Sample moisture trend
Static illustration · does not plot uploaded or NASA data
Soil moisture Sample reference% moisture
4035302520
SAMPLE REFERENCE
6 AM9 AM12 PM3 PM6 PM9 PMSample end
Illustrative chart only. No measured trend or forecast is calculated from this drawing.
Crop conditions
Sample growing profile · not measured
🌱ACTIVE CROP⌄
GROWTH STAGEFlowering
GROWING DAYS54 days
Sample health score87/100
Illustrative profile value · not a diagnosisSample
✦ RULE-BASED PROTOTYPE
Prototype irrigation plan
Calculated from input records · verify present conditions
Monitor Zone ASuggested
Unvalidated sample rule. Recalculate to inspect the prototype result.
Recommendation dismissed. Recalculate to show it again.
Equations and crop thresholds are unvalidated prototype assumptions. Verify field conditions before taking action; this demo does not control irrigation equipment.
What’s influencing the plan
Key factors in this recommendation
Soil moisture31% · below demo targetRule input
Weather demandDemand proxy 4.8 mm · samplePrototype
Input rainfall0 mm · sample; forecast shown separatelyInput
Crop stageFlowering · sample profilePrototype
Rainfall outlook FORECAST
Open-Meteo · today and the next six days · separate from NASA POWER historical weather
Not fetched.
Rain-aware review
Fetch a forecast to review possible rain-related changes.
Daily rainfall = rain + showers, excluding snow. Today is a whole-day estimate, not remaining rainfall. Forecasts are uncertain; no amount is automatically deducted from irrigation. Effective rainfall, infiltration, runoff, root-zone storage and safe timing need local agronomic validation.
Weather data: Open-Meteo · API definitions · CC BY 4.0. Rain + showers are combined here for display.
Soil moisture model TRAINED ML · RESEARCH
24-hour sensor forecast · historical evaluation at Arnesano, Italy · open-field tomato
This model runs in your browser on real held-out farm records. It does not use the selected dashboard crop, coordinates, uploaded CSV or Open-Meteo forecast. It cannot recommend an irrigation amount or predict your field without local validation.
Loading trained model…
Chronological evaluation
Values use the source probe’s 0–100 relative scale, not calibrated volumetric water content. One field and season; sensor faults, missing data and future irrigation limit interpretation. This research result does not change the rule-based irrigation plan.
Data adapted from Adamo et al. (2026), Mendeley Data V2 · CC BY 4.0 · Evaluation details. AISIS filters and aggregates sensor records, trains the model and computes these results.
Bring your field data
Upload validated daily records to update dashboard calculations.
Drop a prepared daily CSV here
CSV files are processed locally in your browser. Required: date (YYYY-MM-DD), increasing dates, and a measurement with explicit units (e.g. soil_moisture_pct). See README for the schema.
Input preview1 sample recordUnvalidated rulesIllustrative sample
PROTOTYPE · DECISION SUPPORT
Built to make each drop count.
This prototype uses illustrative sample profiles, local CSV records and optional NASA POWER historical weather. Dashboard rules and equations are not scientifically validated. Open-Meteo provides separate forecast estimates on request. No live sensor is connected. A separately trained soil-moisture model runs historical test cases in the research panel; it does not control the irrigation plan. The separate Python pipeline prepares experimental features; it does not run on upload.
NASA POWER · weatherKaggle · optional CSVUSDA NRCS · US soil profiles