Greenhouse GFM

The Greenhouse Gap Filling Module calculates greenhouse gas emissions from heated greenhouse cultivation of vegetables. Greenhouse production can increase climate impact by up to 10x compared to field-grown equivalents, making this module critical for accurate environmental assessments of fresh produce.

Quick Reference

Property	Description
Runs on	`ModeledActivityNode` with `FoodProductFlowNode` parent containing matched vegetable terms
Dependencies	`OriginGapFillingWorker`, `AttachFoodTagsGapFillingWorker`, `ConservationGapFillingWorker`
Key Input	Production date, origin country, matched vegetable type
Output	Heating emissions, infrastructure emissions, electricity emissions
Trigger	Product matched to one of 7 greenhouse vegetables

When It Runs

The module triggers when:

A product is matched to a covered vegetable term
The product has a valid origin country with Meteonorm climate data
The product does NOT have an exclusion tag (frozen, canned, dried)
Heating is required (based on climate calculations)

Key Output

The module adds three types of emissions to the calculation graph:

Heating: Based on country climate, vegetable requirements, and production date
Infrastructure: Glass and plastic greenhouse materials
Electricity: Vegetable-specific energy consumption

Scientific Methodology

The greenhouse model calculates emissions using the formula:

Total Emissions = Heating + Infrastructure + Electricity

Each component is calculated per kilogram of vegetable produced.

Heating Demand Model

The heating demand model is based on Stössel et al. (2012) and uses a heat balance approach:

Q_heating = (Q_trans + Q_air) * (T_in - T_out) - Q_solar * G

Where:

Q_trans: Heat transmission through greenhouse envelope
Q_air: Heat loss through air exchange
T_in: Required inside temperature for the vegetable (Kelvin)
T_out: Monthly average outside temperature (from Meteonorm)
Q_solar: Solar heat gain factor
G: Monthly average solar radiation (from Meteonorm)

Heat Transmission Factor (Q_trans)

U_VALUE = 3.4  # W/m2/K - thermal insulation value
CLADDING_AREA = 54_978.2  # m2 - greenhouse envelope area

Q_TRANS_FACTOR = U_VALUE * CLADDING_AREA  # = 186,925.88 W/K

The U-value of 3.4 W/m2/K represents a weighted average for conventional greenhouses according to Scharfy et al. (2017).

Air Exchange Factor (Q_air)

AIR_EXCHANGE_NUMBER = 0.24  # air changes per hour
BUILDING_VOLUME = 259_506.0  # m3
SPEC_VOLUMETRIC_ENERGY_CONSTANT_FOR_AIR = 0.32  # Wh/m3/K

Q_AIR_FACTOR = AIR_EXCHANGE_NUMBER * BUILDING_VOLUME * SPEC_VOLUMETRIC_ENERGY_CONSTANT_FOR_AIR
# = 19,924.46 W/K

Solar Heat Gain Factor (Q_solar)

Q_SOLAR_FACTOR = 0.609 * 46_800.0 * 0.99 * 0.9 * 0.7
# = 17,849.82 W/(W/m2)

Components:

0.609: Solar transmittance of greenhouse cover
46,800 m2: Ground area exposed to sun
0.99: Fraction of solar radiation reaching plants
0.9: Utilization factor
0.7: Conversion efficiency

Heating Emissions Calculation

Heating [kg CO2eq/kg] = Heating demand [MJ/kg] * Heating mixture [kg CO2eq/MJ]

The heating mixture depends on the country and represents the average mix of natural gas, heating oil, and renewables used in greenhouse heating.

Infrastructure Calculation

Infrastructure emissions account for the greenhouse building materials over their 30-year lifespan:

FRACTION_OF_GLASS_IN_BUILDING_MATERIAL = 0.604  # 60.4%
FRACTION_OF_PLASTIC_IN_BUILDING_MATERIAL = 0.396  # 39.6%
AREA_OF_THE_WINDOWS_EXPOSED_TO_THE_SUN = 46_800  # m2

The infrastructure mix (60.4% glass, 39.6% plastic tunnel) is based on Swiss greenhouse statistics from Scharfy et al. (2017).

Calculation:

# m2*year of greenhouse needed per kg of vegetable
building_material_per_kg = greenhouse_area / total_yield_per_year * production_amount

# Split between glass and plastic
glass_material = building_material_per_kg * 0.604
plastic_material = building_material_per_kg * 0.396

Electricity Calculation

Electricity consumption is vegetable-specific and multiplied by the country's electricity mix:

Electricity [kg CO2eq/kg] = Energy input [kWh/kg] * Electricity mix [kg CO2eq/kWh]

The electricity mix uses Ecoinvent's "market for electricity, low voltage" for each country.

Implementation Details

Covered Products

The module runs when a product is matched to one of these FoodEx2 terms:

FoodEx2 Code	Vegetable	Model Type	Description
`B1458`	eggplant	eggplant	Eggplant
`A00JD`	eggplant	eggplant	Aubergines
`A00JM`	cucumber	cucumber	Cucumbers
`A00JR`	cucumber	cucumber	Courgettes (zucchini)
`A00KY`	lettuce	lettuce	Head lettuce
`A00MJ`	lettuce	lettuce	Spinach
`B4946`	lettuce	lettuce	Batavia lettuce
`A00KX`	lettuce	lettuce	Lettuces generic
`A1563`	lettuce	lettuce	Iceberg lettuce
`A0DLB`	lettuce	lettuce	Lettuces and similar
`A1612`	lettuce	lettuce	Oakleaf lettuce
`A00LB`	lettuce	lettuce	Lollo rosso
`A00JA`	bell pepper	bell pepper	Sweet peppers
`A00QV`	radish	radish	Radishes
`A00LM`	radish	radish	Roman rocket and similar
`B2474`	radish	radish	Rocket
`A0DMX`	tomato	tomato	Tomatoes
`A00HY`	vine tomato	vine tomato	Cherry tomatoes

Vegetable Parameters

Each vegetable type has specific cultivation parameters:

Vegetable	Growing Days	Inside Temp (K)	Yield (kg/m2/month)	Electricity (kWh/kg)
eggplant	50	291.15 (18C)	3.15	0.5492
cucumber	32	291.15 (18C)	4.36	0.1982
lettuce	60	281.15 (8C)	1.74	0.4636
bell pepper	41	293.15 (20C)	1.97	0.5746
radish	51	279.15 (6C)	1.36	0.33798
tomato	127	291.15 (18C)	4.66	0.2207
vine tomato	127	291.15 (18C)	4.72	0.2099

Data sources: ProfiCost-Tool (Chollet et al., 2012), Swiss Association of Vegetable Producers technical handbook.

Covered Countries

The module has Meteonorm climate data for 28 European countries:

Code	Country	Code	Country
AT	Austria	IT	Italy
BE	Belgium	LT	Lithuania
BG	Bulgaria	LU	Luxembourg
CH	Switzerland	LV	Latvia
CZ	Czech Republic	NL	Netherlands
DE	Germany	PL	Poland
DK	Denmark	PT	Portugal
EE	Estonia	RO	Romania
ES	Spain	SE	Sweden
FI	Finland	SI	Slovenia
FR	France	SK	Slovakia
GB	Great Britain	TR	Turkey
GR	Greece	HR	Croatia
HU	Hungary	IE	Ireland

Missing EU countries: Cyprus, Malta

Exclusion Rules

The greenhouse model does NOT run if the product has any of these conservation tags:

Term XID	Description
`J0001`	Generic Conserved
`J0136`	Frozen
`J0111`	Canned
`J0116`	Dried

These products are excluded because conserved products don't require greenhouse production at the time of consumption - they were likely produced during the regular growing season.

Note: J0003 (Not Conserved) and J0131 (Cooled) do NOT exclude products from the greenhouse model.

Production Date Handling

The production date determines which months' climate data is used for heating calculations:

If activity_date is specified on the product or any parent node, it's used
The harvest date is calculated as production_date - 3 days
Growing days are backtracked from the harvest date based on the vegetable type
Monthly heating requirements are calculated for each month in the growing period

Example: Tomatoes with production date March 15:

Harvest date: March 12
Growing period: 127 days (October 6 to March 12)
Months with growing days: October (partial), November, December, January, February, March (partial)

Full Code Reference

Heating Calculation

The compute_required_heating method calculates heating demand in MJ per kg of vegetable:

def compute_required_heating(
    self,
    flow_country_code: str,
    flow_processing_date: datetime,
    vegetable_name: str,
) -> float:
    """Compute the necessary amount of heating."""
    days_between_harvest_and_production_date = 3

    harvest_date = flow_processing_date - timedelta(days=days_between_harvest_and_production_date)

    # Get growing days per month based on harvest date and vegetable growth duration
    monthly_growing_days = np.array(days_in_each_month(NUMBER_OF_GROWING_DAYS[vegetable_name], harvest_date))

    # Load country-specific climate data
    monthly_avg_outside_temp = np.array(
        self.gfm_factory.aggregated_meteonorm_data[flow_country_code]["average_temperature"]
    )
    monthly_avg_solar_radiation = np.array(
        self.gfm_factory.aggregated_meteonorm_data[flow_country_code]["average_solar_radiation"]
    )

    # Calculate heat balance
    monthly_temp_difference = REQUIRED_INSIDE_TEMPERATURE[vegetable_name] - monthly_avg_outside_temp
    monthly_required_heating_power_watts = (
        Q_TRANS_FACTOR + Q_AIR_FACTOR
    ) * monthly_temp_difference - Q_SOLAR_FACTOR * monthly_avg_solar_radiation

    # No negative heating (no cooling modeled)
    monthly_required_heating_power_watts = np.array(
        [val if val >= 0.0 else 0.0 for val in monthly_required_heating_power_watts]
    )

    # Convert power to energy
    monthly_required_heating_energy_wh = monthly_required_heating_power_watts * monthly_growing_days * 24
    monthly_required_heating_energy_kwh = monthly_required_heating_energy_wh / 1000

    # Calculate yield during growing period
    average_yield_kg_per_day = (
        AVG_YIELD_PER_M2_PER_MONTH[vegetable_name] * AREA_OF_THE_WINDOWS_EXPOSED_TO_THE_SUN * 12 / 365
    )
    yield_in_growing_days_kg = average_yield_kg_per_day * sum(monthly_growing_days)

    # Convert to MJ per kg
    mj_in_kwh = 3.6
    monthly_required_heating_energy_mj = monthly_required_heating_energy_kwh * mj_in_kwh

    required_heating_energy_mj_for_kg_of_vegetable = (
        np.sum(monthly_required_heating_energy_mj) / yield_in_growing_days_kg
    )

    return required_heating_energy_mj_for_kg_of_vegetable

Infrastructure Accounting

async def account_for_infrastructure(
    self, calc_graph: CalcGraph, greenhouse_activity: FoodProcessingActivityNode, vegetable_name: str
) -> None:
    """Account for infrastructure of greenhouse."""
    size_of_greenhouse_in_m2 = AREA_OF_THE_WINDOWS_EXPOSED_TO_THE_SUN
    average_yield_in_kg_per_m2_and_month = AVG_YIELD_PER_M2_PER_MONTH[vegetable_name]
    total_yield_per_year = average_yield_in_kg_per_m2_and_month * 12 * size_of_greenhouse_in_m2

    # Calculate m2*year of building material per kg of crop
    area_and_time_of_greenhouse_building_material = (
        size_of_greenhouse_in_m2 / total_yield_per_year * self.node.production_amount.value
    )

    # Split between glass (60.4%) and plastic (39.6%)
    glass_building_material = (
        area_and_time_of_greenhouse_building_material * FRACTION_OF_GLASS_IN_BUILDING_MATERIAL
    )
    plastic_building_material = (
        area_and_time_of_greenhouse_building_material * FRACTION_OF_PLASTIC_IN_BUILDING_MATERIAL
    )

    # Add glass and plastic flow nodes to the graph
    # Connected to Ecoinvent processes for greenhouse infrastructure

Electricity Accounting

async def account_for_electrical_energy(
    self,
    calc_graph: CalcGraph,
    greenhouse_activity: FoodProcessingActivityNode,
    country_code: str,
    vegetable_name: str,
) -> None:
    """Account for electrical energy consumption of greenhouse."""
    energy_needed_per_kg = REQUIRED_ELECTRICITY_KWH_PER_KG[vegetable_name]

    # Create electricity consumption flow
    electricity_consumption_flow = FlowNode(
        uid=UuidStr(uuid4()),
        amount=QuantityProp(
            value=energy_needed_per_kg * self.node.production_amount.value,
            unit_term_uid=self.gfm_factory.kWh_term.uid,
        ),
    )

    # Connect to country-specific electricity market from cache
    electrical_energy_node = self.gfm_factory.electricity_markets_cache.get(country_code)

Data Sources

Meteonorm Climate Data

Climate data is sourced from Meteonorm software (2016 data):

Temperature: Monthly average outside temperature in Kelvin
Solar radiation: Monthly average global solar irradiation in W/m2

The data is stored in the GFM cache and loaded at service startup:

meteonorm_average_temperature_{country_code}: Array of 12 monthly values
meteonorm_average_solar_radiation_{country_code}: Array of 12 monthly values

Heating Mixes

Country-specific heating mixes for greenhouse heating:

Country	Brightway XID	Composition
CH (Switzerland)	`EDB_4220cbaabca343c09b19415d5ab4079f_copy1`	41.3% natural gas, 35.2% heating oil, 23.5% renewables
DE (Germany)	`EDB_4220cbaabca343c09b19415d5ab4079f_copy2`	28% coal, 21% natural gas, 15% fuel oil, 20% renewables, 16% other
NL (Netherlands)	`EDB_4220cbaabca343c09b19415d5ab4079f`	86.2% natural gas, 0.1% heating oil, 13.7% renewables
EU other	`EDB_4220cbaabca343c09b19415d5ab4079f_copy3`	54% natural gas, 46% heating oil

Infrastructure Materials

Ecoinvent processes for greenhouse building materials:

Material	Ecoinvent ID	Unit
Glass greenhouse	`ecoinvent 3.6 cutoff_0ce659c3cfd443a38761058ee62e3f10`	m2*year
Plastic tunnel	`ecoinvent 3.6 cutoff_fddfe51c6959f41ac044089c3a892af7`	m2*year

Calculation Example

Scenario: 1 kg of tomatoes, produced in Sweden on March 15

Step 1: Determine Growing Period

Production date: March 15
Harvest date: March 12 (3 days before production)
Growing duration: 127 days
Start date: November 5 (previous year)

Step 2: Calculate Monthly Growing Days

Month	Days
November	25
December	31
January	31
February	28
March	12
Total	127

Step 3: Calculate Heating Demand

Using Swedish climate data and tomato requirements (18C inside temperature):

Monthly temperature differences calculated
Solar radiation offset applied
Negative values clamped to zero

Result: ~32.3 MJ/kg heating demand

Step 4: Calculate Infrastructure

Total yield/year = 4.66 kg/m2/month * 12 * 46,800 m2 = 2,616,576 kg/year
Building material per kg = 46,800 / 2,616,576 = 0.0179 m2*year/kg

Glass: 0.0179 * 0.604 = 0.0108 m2*year
Plastic: 0.0179 * 0.396 = 0.0071 m2*year

Step 5: Calculate Electricity

Electricity = 0.2207 kWh/kg * Swedish electricity mix emissions

Final Output

The module adds three flow nodes to the calculation graph:

Heating flow (32.3 MJ) connected to Swedish heating mix process
Glass infrastructure flow (0.0108 m2*year) connected to Ecoinvent glass process
Plastic infrastructure flow (0.0071 m2*year) connected to Ecoinvent plastic process
Electricity flow (0.2207 kWh) connected to Swedish electricity market

Known Limitations

Geographic Coverage

Only 28 European countries have Meteonorm data
No coverage for US, Canada, or other major producing regions
No sub-country regional differentiation (important for large countries)

Crop Coverage

Only 7 vegetable types are modeled (tomato, cucumber, lettuce, bell pepper, eggplant, radish, vine tomato)
Herbs, chilis, strawberries, and other greenhouse crops are not covered
Some crops like lamb's lettuce have data available but aren't implemented

Model Assumptions

No heating required = no greenhouse assumed (may underestimate emissions for unheated greenhouses)
Same infrastructure mix used for all countries (based on Swiss data)
Organic production not differentiated (minimal impact difference found)
No country-specific greenhouse growing seasons enforced

Data Age

Meteonorm climate data is from 2016
Heating mixes may not reflect current energy transitions

References

Stössel, F., Juraske, R., Pfister, S., & Hellweg, S. (2012). Life cycle inventory and carbon and water footprint of fruits and vegetables: application to a Swiss retailer. The International Journal of Life Cycle Assessment, 17(9), 1191-1202.
Eymann, L., et al. (2014). Gewächshausmodell v1.0. ZHAW Report for Eaternity.
Scharfy, D., et al. (2017). OFP Report - New Data ZHAW v1.3. Organic Food Print project report.
Chollet, D., et al. (2012). ProfiCost-Tool. Swiss Association of Vegetable Producers.
Meteonorm Software. https://meteonorm.com/

Quick Reference​

When It Runs​

Key Output​

Scientific Methodology​

Heating Demand Model​

Heat Transmission Factor (Q_trans)​

Air Exchange Factor (Q_air)​

Solar Heat Gain Factor (Q_solar)​

Heating Emissions Calculation​

Infrastructure Calculation​

Electricity Calculation​

Implementation Details​

Covered Products​

Vegetable Parameters​

Covered Countries​

Exclusion Rules​

Production Date Handling​

Full Code Reference​

Heating Calculation​

Infrastructure Accounting​

Electricity Accounting​

Data Sources​

Meteonorm Climate Data​

Heating Mixes​

Infrastructure Materials​

Calculation Example​

Step 1: Determine Growing Period​

Step 2: Calculate Monthly Growing Days​

Step 3: Calculate Heating Demand​

Step 4: Calculate Infrastructure​

Step 5: Calculate Electricity​

Final Output​

Known Limitations​

Geographic Coverage​

Crop Coverage​

Model Assumptions​

Data Age​

References​