Previous PageTable Of Contents

Using Life Cycle Assessment (LCA) to assess water use in tomato production

Girija Page1, Bill Bellotti1 and Brad Ridoutt2

1 University of Western Sydney, School of Natural Sciences, Locked Bag 1797, Penrith South DC 1797, NSW, Australia. Email
CSIRO Sustainable Agriculture National Research Flagship, Private Bag 10, Clayton, Victoria. Email


In response to concerns about global water scarcity and food security, water footprints have emerged as one important indicator of sustainable agriculture. In this paper the water footprints of two greenhouse tomato production systems are presented which form part of a wider study on sustainable food production using Life Cycle Assessment (LCA). The water footprints, which offer a quantitative measure of the way a production system contributes to the problem of physical water scarcity, were calculated using a recently developed LCA-based calculation method, taking into account the local water stress where production occurred. For greenhouses located in Sydney and Guyra (NSW northern tablelands), the world normalised water footprints of market tomatoes were 35 and 3 L H2Oe per kg fresh weight respectively. In comparison, the water-use efficiency of these systems was 50 L (Sydney) and 39 L (Guyra) per kg. Although water-use efficiency is popular amongst agronomists and is important for benchmarking local resource-use efficiency, it does not address the wider question of sustainable freshwater use. Water footprint on the other hand is capable of sending signals on the way the agricultural production system limits the availability of freshwater for the environment and/or other human uses. Metrics such as water footprint offer a useful and additional perspective for moving towards environmental sustainability.

Key Words

Water-use efficiency (WUE), water footprint, tomato, water scarcity, sustainability


There is growing awareness in the public and policy arenas that the rate of withdrawal of freshwater to meet human production and consumption is not sustainable and must reduce in intensity. Freshwater scarcity is increasing in many parts of the world and is recognised as a major constraint on current and future food production and a threat to the health of freshwater ecosystems (Rockstrom et al. 2009). Likewise, agriculture is a major consumer of freshwater in the Australian economy, accounting for over 65% of the total water consumption, 91% of which is for the irrigation of crops and pastures (ACIAR 2010). Demand for freshwater by agriculture to meet food production of an increasing world population is clearly in competition with water for the environment.

In response to increasing awareness of freshwater sustainability, interest in water footprinting is developing as a way to make transparent both direct and indirect water use associated with agriculture. Most of the approaches to sustainable vegetable production consider the “crop per drop”: yield/evapotranspiration (ET) or yield/irrigation ratio as measures of eco-efficiency, sometimes called water-use efficiency by agronomists. Another approach considers the virtual water content (VWC) which is similar to the water use efficiency in many ways. Researchers calculating VWC of crops have tended to be concerned about the flows of virtual water in relation to trade (Chapagain and Orr 2009). Although, these approaches are useful to account for the absolute amounts of water used, the need to consider the regional impact factors within LCA has been pinpointed (Pfister et al. 2009a). Considering the regional impact factors will enable the estimation of the potential social or environmental harm arising from physical water scarcity as a result of the consumptive water use (Ridoutt and Pfister 2010). In this study, we apply a recently developed LCA-based methodology to assess the environmental impact of water use in food production. This method is based on the premise that the potential of the human activity to cause local water scarcity needs to be considered as a key environmental impact (Ridoutt and Pfister 2010). We illustrate the method using two case studies of tomato production. We have chosen tomato as a case study because after potato, it dominates the production and trade amongst vegetables. We compare the findings of the LCA-based approach with another form of water-use measurement in agriculture and horticulture, namely the water-use efficiency.


Greenhouse production systems

Inventory on the key biophysical inputs and outputs for a typical production year was gathered from four hydroponic greenhouse tomato growers in New South Wales (Table 1): three from Sydney and one from Guyra (410 km north of Sydney). The greenhouses in Sydney were considered ‘med-tech’, meaning that they had partial control over the growing conditions and had moderate use of inputs compared to the greenhouse in Guyra which was ‘hi-tech’, meaning it had complete control over the growing conditions with high intensity of inputs. Information was gathered through semi-structured interviews with the growers. The studied systems included processes up to the greenhouse gate.

The typical size of the med-tech greenhouse is 7000 m2. The tomato crop is grown using an ‘open hydroponic system’, meaning that the run-off from the greenhouse is allowed to drain in the surrounding paddocks. The hi-tech greenhouse is a modern growing structure offering superior crop performance. Environmental controls are completely automated. The total size of the hi-tech greenhouse is 20 ha, having 5 ha under one roof. The runoff from the hi-tech system (30% of applied water) is recycled into the greenhouse. Rainwater harvested from the greenhouse satisfies 30% of the water requirement for tomato production; the excess rainwater is stored in a pond and drains into the environment. The remaining 40% of the water requirement is supplied through irrigation.

The information on key biophysical inputs/outputs of the med-tech and hi-tech greenhouse tomato production systems over a typical production year is presented in Table 1.

Table 1. Description of key information of the two tomato greenhouse production systems over a typical production year.




Yield kg/ha



Planting density plants/m2



Active pesticides kg/ha



Calcium nitrate kg/ha



Potassium nitrate kg/ha



Monopotassium phosphate kg/ha



Potassium sulphate kg/ha



Magnesium sulphate kg/ha



Phosphoric acid (L/ha)



Diesel L/ha



Coal t/ha



Butane t/ha



Water requirement for tomato cultivation* m3/ha



*includes water requirement for tomato cultivation, washing sheds and cooling the greenhouse in summer

Note: Although the fertiliser requirement may appear high, the N content does not constitute over 0.6% of the total fresh weight of tomatoes.

Water-footprint calculation method at the product level

Water footprints were calculated following the method of Ridoutt and Pfister (2010). Accordingly, the main concern relating to water consumption in agri-food product life cycles is the potential to contribute to water scarcity, thereby limiting the availability of freshwater for human uses and for the environment. The inventory stage incorporated consumption of blue-water resources (irrigation water from surface and groundwater sources), and the change in blue-water availability arising from land use (i.e through altered drainage and runoff from greenhouse cultivation). In addition to greenhouse irrigation, blue water is also used indirectly during the extraction of raw material and energy sources such as coal or natural gas and for the manufacture of farm inputs (eg. fuel, fertiliser). The Australian LCA database, implemented in Simapro version 7.2.2, was used for estimating the blue-water use from these processes; international databases were used when relevant information was not available from the Australian database. Blue water associated with the production of capital goods, such as greenhouses, office buildings and machinery, was excluded from the assessment. The water footprint was expressed for the production of one kg of fresh tomato at the greenhouse gate.

The consumption of green water (rainfall) per se does not contribute to water scarcity. Until it becomes blue water, green water does not contribute to environmental flows, nor is it accessible for other human uses beyond the immediate property (Ridoutt and Pfister 2010). However, land use affects the partitioning of green water to blue water. In the case of tomato production, blue-water availability is affected by the construction of greenhouses and by the local collection and use of rainwater. The baseline situation (no greenhouse) was modelled using the generalized equation of Zhang et al. (2001) relating ET to precipitation (P) for grassed catchments (equation 1). The difference between P and ET was assumed to become blue water as deep drainage and stream flow. The model was then re-run taking into account the collection of storm water in farm ponds, return to greenhouse as irrigation and losses via evaporation. This enabled the estimation of the change in blue-water availability as a result of greenhouse tomato production.


It is rational to consider the impacts of nutrient runoff (grey water) (Chapagain and Orr 2009) from the greenhouse on freshwater availability. However, grey water was not considered in this research. The runoff from the med-tech greenhouse is drained in the surrounding paddock. Since the emissions from the paddock are outside the system boundary they were not included in the inventory. For impact assessment, local characterisation factors for water scarcity for Sydney (0.397) and Guyra (0.015) were taken from the Water Stress Index (WSI) of Pfister et al. (2009b). The WSI takes into consideration parameters such as watershed withdrawals to availability ratio, degree of regulation of flows and variability in precipitation (Pfister et al. 2009b). The average Australian WSI was used in relation to blue-water consumption for the extraction of raw material and production of farm inputs (0.402). To estimate the product water footprint, each instance of blue-water consumption was multiplied by the relevant WSI and then summed across the system. The product water footprint was normalized by dividing by the global average WSI (0.602) and expressed in the units H2Oe (equivalent). This manner of expressing product water footprints has been found to be useful as it enables a decision maker to quantitatively compare the pressure exerted on freshwater systems through the consumption of a product (i.e. via indirect freshwater consumption) with an equivalent volume of direct freshwater use.

Results and Discussion

The production of one kg of tomato at the Sydney greenhouse gate required 53 L of blue water (Table 2), 95% occurring at the greenhouse and the balance associated with the production of farm inputs. The Sydney greenhouse had no negative effect on the blue-water availability, since the rain falling on the greenhouse drained into the surrounding pasture. The total blue-water requirement for the production of a kg of fresh tomato at the greenhouse located in Guyra was 22 L, 68% occurring at the greenhouse, 18% associated with the production of farm inputs and the balance related to rainwater harvesting.

Table 2. Water use metrics for one kg of fresh tomato produced in Sydney and Guyra.




Total blue water consumption1 (L)



Water footprint2 (L)



World equivalent water footprint3 (H2Oe)



Water use efficiency4 (L)



1 Blue water used directly for tomato cultivation + blue water used indirectly in the extraction and production of inputs + change in blue water availability as a result of greenhouse production
Blue water multiplied by respective WSI and then summed across the system
Water footprint value divided by global average WSI of 0.602
Water requirement for tomato cultivation (from Table 1) (L) divided by tomato yield (kg)

Differences in water footprints for a kg of tomato produced in Sydney and Guyra can be explained by the water-stress characterisation factors associated with the two locations. Following Pfister et al. (2009b), the water-stress index for the med-tech greenhouse in the Sydney area is about twenty-six times higher than the water-stress index for the hi-tech system. Taking yields, consumptive water use, local WSI and other inputs into consideration, a kg of tomatoes produced at the Sydney greenhouse had a water footprint of 21 L, while that at the Guyra greenhouse had 2 L. This emphasises the importance of including the water-stress characterisation factors relevant to each location where water is consumed (Ridoutt and Pfister 2010). The normalised water footprints suggest that a kg of tomato grown in Sydney and in Guyra had an equivalent potential to contribute to freshwater scarcity (as defined by the WSI, Pfister et al. 2009a) as the direct consumption of 35 and 3 litres of water globally (at the world average WSI of 0.602). However, locating the greenhouse further away from the consumer in Sydney (in this case Guyra) will likely necessitate higher fossil-fuel use to transport the tomatoes to market (i.e. the ‘food-mile’ issue). The overall sustainability of alternative tomato production systems, taking climate impacts, water use and other key environmental impact categories into account, will be assessed as part of the next steps in this project.

The water-use efficiency was 50 L/kg in Sydney and 39 L/kg in Guyra (Table 2). Although water-use efficiency has been popular among agronomists as a measure of relating crop yield to growing-season rainfall/irrigation, it can neither be related to the potential to cause local water scarcity at the production site, nor does it differentiate the type of water (blue, green, grey) used. This limits the use of water-use efficiency for guiding changes required to move towards long-term sustainability of water use. These results therefore suggest caution in the public communication of water-use efficiency or other similar indicators (e.g. virtual water content) which consider volumes only, as a lower value may not indicate more sustainable freshwater use. In their effort to address issues of sustainability, Walmart, the world’s largest retailer has already indicated the development of indicators for measuring the water-use efficiency of their products ( In the short term, it would be good to see retailers requiring growers to report on water-use efficiency as this will encourage the growers to be mindful of their water use in order to seek best practice for their growing region. However, a grower might be water efficient, but located in a region that uses water unsustainably with serious consequences for the environment. Hence, water-use efficiency as a sole indicator of water use can be misleading.


The water footprints as estimated in this research provided a meaningful comparison between greenhouse tomato production systems at different locations, in their potential to contribute to freshwater scarcity. Having both water-use efficiency and water footprint as indicators of water use is helpful. This recognises the ability to achieve short-term benefits (through water use efficiency), while also recognising that longer-term structural change is needed to move towards environmental sustainability (through water footprints). The results suggest that site selection for greenhouse establishment is important for reducing the water footprint. Relocation of the greenhouse industry in Sydney to other areas that have a lower water-stress index can help alleviate water scarcity. However, this implies the movement of the production site further away from the consumer (Sydney being a major consumer of fresh tomatoes) and with resultant other environmental impacts such as an increase in fossil-fuel use and associated greenhouse gas emissions from ‘food-miles’. These trade-offs will be explored in the ongoing research by considering water footprinting along with other indicators of environmental sustainability such as energy use and global warming.


ACIAR (2010). Reducing water use in agriculture. Accessed 6 March 2010

Chapagain AK and Orr S (2009). An improved water footprint methodology linking global consumption to local water resources: A case of Spanish tomatoes. Journal of Environmental Management, 90, 1219-1228.

Pfister S, Stoessel F, Juraske R and Koehler ASH (2009a). Regionalized LCIA of vegetable and fruit production: Quantifying the environmental impacts of freshwater use. Paper presented at the International Conference on LCA in the Agri-Food Sector - Towards a sustainable management of the Food chain. ISBN 978-3-905733-10-5, available at:

Pfister S, Koehler A and Hellweg S (2009b). Assessing the Environmental Impacts of Freshwater Consumption in LCA. Environmental Science & Technology, 43, 4098-4104.

Ridoutt B G and Pfister S (2010). A revised approach to water footprinting to make transparent the impacts of consumption and production on global freshwater scarcity. Global Environmental Change, 20, 113-120.

Rockstrom J, Steffen W, Noone K, Persson A, Chapin FS, Lambin EF et al. (2009). A safe operating space for humanity. Nature, 461, 472-475.

Zhang L, Dawes WR and Walker GR (2001). Response of mean annual evapotranspiration to vegetation changes at catchment scale. Water Resources Research, 37, 701-708.

Previous PageTop Of Page