Field and spatial
Field boundary, hectares, GPS, management zones, elevation, slope/aspect, erosion class, drainage, compaction, rooting depth, previous yield maps and sampling design.
RAVAM working model · v2.5
Balkan agricultural intelligence map
A district-level screening layer for where to investigate nitrogen efficiency, phosphorus recovery, liming and micronutrients. It is not a field prescription: every recommendation becomes valid only after parcel soil analysis, crop, growth stage and yield target.
Dominant first agronomic question
Serbia · select one of 25 districts
Coverage and data sources
Coverage note: Kosovo is presented on its own tab through five statistical areas following the Statistical Office of the Republic of Serbia classification. Labels and boundaries are used for agricultural screening only and do not determine political status. Serbia and Kosovo geometry is adapted from TUBS/Wikimedia Commons, CC BY-SA 3.0. Hungary county geometry is adapted from the public-domain HU counties blank map; crop framing is cross-checked against Hungarian Central Statistical Office county data and Nébih soil guidance. Montenegro municipal geometry is adapted from the current 2022 Wikimedia Commons administrative map, CC BY-SA 3.0 DE; agronomic framing uses the Montenegro Ministry soil-analysis guidance, its livestock overview and viticulture guidance.
Bosnia and Herzegovina coverage: The selectable layer uses 10 Federation cantons, six Republika Srpska planning macroregions and Brčko District. Geometry is adapted from the April 2026 Wikimedia Commons municipalities-and-regions map, CC BY-SA 4.0; crop framing is cross-checked against the Federation agricultural production bulletins and the Republika Srpska municipal statistics. Administrative and agronomic boundaries are indicative screening layers, not parcel evidence.
North Macedonia coverage: The selectable layer follows all eight official statistical regions and 2024 regional crop data in the State Statistical Office Regions of North Macedonia 2025. Region geometry comes from geoBoundaries (CC BY 4.0; Runfola et al., 2020), whose source metadata cites EuroGeographics, NTES and the State Statistical Office. Regional crop and soil framing is for screening; parcel sampling and local agronomic validation remain essential.
Croatia coverage: The selectable layer follows 20 counties plus the City of Zagreb. Crop framing uses the Croatian Bureau of Statistics agriculture programme, including its county-level Agricultural Census 2020 and 2024 crop releases. Geometry comes from geoBoundaries; current metadata identifies 21 units and credits OpenStreetMap/Wambacher under ODbL 1.0. County profiles are screening hypotheses, not parcel prescriptions.
Romania coverage: The selectable layer follows all 41 counties plus Bucharest. Crop framing uses the Romanian National Institute of Statistics agriculture programme, its 2023 county structural survey and 2024 crop production release. Geometry comes from geoBoundaries (42 ADM1 units, World Bank source, CC BY 4.0). The eight regional belts and county profiles are screening hypotheses, not parcel prescriptions.
Moldova coverage: The selectable layer follows 37 first-level administrative areas: 32 districts, Găgăuzia, Chișinău, Bălți, Bender and Transnistria. Geometry comes from geoBoundaries (2020 representation; UNHCR/OCHA source, CC BY 3.0 IGO). Crop framing uses the National Bureau of Statistics agriculture programme and its regional statistics. Boundaries and the five agronomic belts are screening layers only and do not determine political status or replace parcel analysis.
Albania coverage: The selectable layer follows all 12 counties. Crop framing uses INSTAT’s Regional Statistical Yearbook 2024 and its regional statistics programme. Geometry comes from geoBoundaries (12 ADM1 units; 2021 geoBoundaries/Wikipedia source, Public Domain). The six belts and county profiles are screening hypotheses, not parcel prescriptions.
Bulgaria coverage: The selectable layer follows all 28 provinces. Crop framing uses the Bulgarian National Statistical Institute agriculture programme, the Ministry of Agriculture’s crop statistics and Agricultural Census 2020. Geometry comes from geoBoundaries (28 ADM1 units; 2019 geoBoundaries/Wikimedia source, Public Domain). The nine belts and province profiles are screening hypotheses, not parcel prescriptions.
Slovenia coverage: The selectable layer follows all 12 statistical regions using the 2024 Eurostat GISCO NUTS 3 geometry. Crop framing is cross-checked against the Statistical Office of Slovenia 2024 crop-production release by statistical region. Regional belts are screening hypotheses; parcel soil, crop and tissue evidence remains decisive.
Greece coverage: The selectable layer follows all 13 NUTS 2 regions using 2024 Eurostat GISCO geometry. Crop framing uses ELSTAT crop area and production by Region and the 2023 Annual Agricultural Statistical Survey. Drought, water quality, salinity and perennial-crop profiles require parcel-level verification.
European Turkey coverage: The selectable layer uses the five province-level areas Edirne, Kırklareli, Tekirdağ, İstanbul and Çanakkale. Geometry comes from geoBoundaries; crop framing uses the Turkish Ministry of Agriculture and Forestry’s Meriç–Ergene basin description and Thrace cropland erosion research. Full İstanbul and Çanakkale province boundaries are shown even though they straddle Europe and Asia; this is a screening layer, not a parcel prescription.
Nutrient-budget decision support · v1.0
Advanced fertilizer prescription calculator
Build an explainable N–P₂O₅–K₂O budget from crop demand, measured soil supply, nutrient credits and expected field recovery. The result is a planning estimate for agronomist review—not a laboratory interpretation or legal prescription.
Input confidence
0%
Add field-specific evidence to improve the estimate.
1Field and crop Essential
Use a realistic yield supported by field history, water supply and management—not regional maximum yield.
2Crop demand coefficients Verify locally
Coefficients must represent the chosen budgeting basis—total uptake or harvested removal—and match the local recommendation system. Do not mix bases.
3Soil sampling and laboratory supply Essential
Raw P and K laboratory values cannot be universally converted to kg/ha. Enter calibrated supply only from a local method, crop response calibration and agronomist interpretation. For annual N, use parcel- and year-specific N-min where appropriate.
4Nutrient credits and organic sources Avoid double counting
5Water, weather and expected recovery 4R management
Recovery efficiency is not a universal constant. Set it from local trials and field conditions; poor timing, placement, waterlogging, drought or extreme pH can lower it.
6Plant feedback and diagnosis In-season refinement
7Geology, contamination and environmental constraints Safety gate
Editable fertilizer materials
Approved nutrient target
Use this mode only when an agronomist or existing recommendation has already established the mineral fertilizer target.
Full professional input register — what the best possible model may need
Crop and yield
Species, cultivar/rootstock, planting density/date, phenology, marketable vs total biomass, realistic yield and quality target, residue fate, rotation and rooting pattern.
Soil physical
Texture by depth, bulk density, stones, available water capacity, infiltration, hydraulic conductivity, depth restrictions, aggregate stability and temperature/moisture.
Soil chemistry
pH and buffer pH, organic C/humus, CEC/base saturation, EC/sodicity, carbonates, mineral N by depth, calibrated P/K, S, Ca, Mg and method-specific interpretation classes.
Micronutrients and safety
B, Zn, Fe, Mn, Cu, Mo, Cl and Ni where relevant; Cd, Pb, As, Hg, Cr, Ni and other contaminants; radiological screening only where geology/history justifies it.
All nutrient sources
Fertilizer carryover, manure/slurry/compost analyses and availability, legumes, residues, biosolids or by-products, irrigation water, atmospheric deposition and biological fixation.
Water and weather
Rainfall/irrigation amounts and timing, evapotranspiration, drought/waterlogging risk, nitrate and salts in water, application forecast, leaching/runoff/volatilization risk.
Plant feedback
Tissue tests by stage and plant part, SPAD/reference strips, canopy imagery, NDVI, deficiency symptoms, disease/pest pressure, lodging and harvest nutrient removal.
Application engineering
Material grade and assay, density, granule-size compatibility, hygroscopicity, blend segregation, spreader calibration, bout width, placement, timing, splits and incorporation.
Environment and law
Water-protection zones, setbacks, vulnerable areas, local N/manure limits, frozen/saturated soil restrictions, slope and erosion controls, records and product registration.
Economics and inventory
Product delivered cost, operations, storage, labor, yield response probability, crop value, quality premium, risk tolerance and stock constraints—kept separate from agronomic safety limits.
Verification and learning
Untreated/reference strips, replicated field trials, as-applied maps, rainfall log, tissue checks, yield/quality result, nutrient balance and post-season model calibration.
Calculation: crop demand = yield target × crop coefficient + adjustment. Available supply = measured soil supply + mineralization/rotation/residue credits + first-year organic availability + irrigation credits. Mineral target = max(0, (demand − available supply) ÷ expected recovery). The target is then allocated to the editable materials. The calculator uses P₂O₅ and K₂O fertilizer-label conventions. Lime, S, Ca, Mg, micronutrients and contaminants require separate calibrated modules and are intentionally not converted into a dose here.
Method framework: USDA NRCS Nutrient Management 590 · FAO Fertilizer and Plant Nutrition Guide · Serbian N-min field guidance. Local calibration, current Serbian rules and accredited laboratory interpretation control the final recommendation.
Measure the parcel
Soil sampling, pH, organic matter, available nutrients, texture, drainage and—where needed—leaf analysis.
Define the crop target
Crop, growth stage, expected yield, previous crop, manure credits, irrigation and weather risk determine the nutrient requirement.
Calculate and verify
Convert the confirmed nutrient target to product quantities, apply at the correct timing, then verify crop response and update the field record.