4.1 Cultivation
Over the 10year period (from 2005 to 2015), based on the data, we observed a range of changing cultivation practices. Diesel use and water use decreased by 23% and 45%, respectively, per unit of harvested tomato. On-farm reductions in water use are tied to reductions in energy and fuel use (used to pump irrigation water) and a shift in irrigation technologies such as drip irrigation. Fifty percent of surveyed growers shifted from furrow to drip irrigation between 2005 and 2015. Among those who did not switch, 13% continued to use furrow irrigation in 2015, and 13% were already using drip irrigation in 2005 (and 2015). The trend towards drip irrigation reflects a broader trend in California, with drip increasing by ~ 38%, while furrow irrigation decreased by ~ 37% for vegetable, orchard, and vineyard crops. Low-volume irrigation technologies increase flexibility in irrigation scheduling and improve control of moisture levels (Tindula et al.
2013), and, in the case of tomatoes, increase crop yields (Geisseler and Horwath
2014).
Surveyed growers also made a shift in fertilizer choice; for example, the data show a 25% reduction in the use of 4-10-10 (NPK) and a 10% increase in the use of calcium ammonium nitrate (CAN) 17. Notably, the data collected show there is no clear correlation between water and fertilizer inputs, and yield increases—from 92 to 123 metric tons per hectare (41 to 55 US tons per acre) in 2005 and 2015 respectively—among surveyed growers. In general, while large differences in soil type and thus water holding capacity may account for some of the differences in input rates, these results also suggest that there may be opportunities for some producers to improve water and fertilizer use efficiencies.
Ultimately, the upstream impacts need to be balanced with the downstream impacts that occur after field application of the product. A review by Rosenstock et al. (
2016) shows that different types of N fertilizers result in different NH
3-N and N
2O emissions after application. For example, although this study has found that urea-based fertilizers (such as UN32) have lower emissions in their production compared with some blends, Rosenstock et al. (
2016) show that they can result in as much as 40% higher NH
3-N emissions in the field than mixed ammonia and nitrate fertilizers, such as CAN17. On the other hand, ammonia products (such as aqua ammonia) can lead to 40–60% higher N
2O emissions in the field than urea-based or sulfate fertilizers.
4.2 Processing facility
The results show processing facility practices decreased impacts per functional unit over time (from 2005 to 2015) at the facility processing stage. Natural gas-related effects decreased by 27% and 5% for diced and paste products, respectively. Processing facility grid electricity and water use also reduced by 27% and 5% and 22% and 5% for diced and paste products, respectively. Because grid electricity generation entails significant amounts of freshwater use, decreases in electricity use translate to reductions in life cycle freshwater use.
Although we observed similar general trends in the 2005 and 2015 and 2010 and 2015 data (e.g., for reductions in natural gas use), our observations of the data indicate that evaluation of the between-facility variability requires process or equipment-level data, which is beyond the scope of this project. As such, an only general discussion about the cause of variability is possible. For example, thermal processes, such as evaporation, tend to be more energy-intensive per ton of product than mechanical methods, such as dicing. Energy source (e.g., grid electricity vs. onsite solar energy generation) and the rate of energy use (quantity of energy used per defined timeframe) and water demands vary depending on the product specs (e.g., select color, titratable acidity, and sugar content), which affect process-level equipment and settings (e.g., temperature) applied to produce that product.
Regarding the inter-annual variability, based on feedback from survey respondents, this variability is, in part, a reflection of the tomato throughput for the season. Basically, once the facility is made operational for the season, much of the equipment is kept running regardless of occasional gaps in tomato input. Thus, in high-yield or high-throughput years, the whole facility is more efficient than in low-yield or low-throughput years.
4.3 Comparison with other studies
In the current study, top impacts in greenhouse operations included fuel (natural gas use) for heating and electricity for pumps, lighting, etc. The second most impactful inputs include growth medium materials (vermiculite and peat), contributing to GWP, TPE, and FWC, as well as the AP, POCP, ODP, and EP impacts. For similar impact categories (AP, EP, and POCP), Del Borghi et al. (
2014) show that packaging is a leading contributor to POCP (up to ~ 63% of the total impacts depending on the product and its packaging), and cultivation is the top contributor to the EP impacts. Some previous studies (e.g., Dias et al.
2017; De Marco et al.
2018) have examined greenhouses using LCA, and another essential factor in determining impacts at this phase in the supply chain is whether the heating is required. In general, because California has a Mediterranean climate, it does not need much heating relative to some of the other regions of study like Canada and Germany.
Results of this LCA for the cultivation stage align with those from previous LCAs of tomatoes. These results include the significant contribution of fossil fuels (diesel) (Del Borghi et al.
2014; De Marco et al.
2018), electricity use for irrigation (De Marco et al.
2018; Ntinas et al.
2017), and fertilizer production and use (Ntinas et al.
2017). The primary data for inputs in the cultivation phase in the current study are similar to those reported in Del Borghi et al. (
2014), but the GWP values reported in the current study are on the lower end of the GWP values reported there, ranging from 0.4 to 0.6 kg CO
2e per kg cultivated tomato. The modeled inputs reported in De Marco et al. (
2018) are an order of magnitude higher per unit of the final product (e.g., for diesel inputs at the cultivation phase) compared with the current study, resulting in higher reported GWP values in their research. Overall, these studies show that across the supply chain, cultivation is the main contributor to ecotoxicity and human health impacts, including ozone depletion potential (based on CML methodology). Differences between the results of these studies may be due to weight ratios. For example, a wide range of products assessed in the Del Borghi study, including tomato purée, chopped tomatoes, and peeled tomatoes in tomato juice, along with mashed tomato in the De Marco study, do not report weight ratios for all products. Compared with the Brodt et al. (
2013) study with the same weight ratios, water impacts are within range of the values observed in the current study ~ 130 to 610 kg water per kg product, for diced and paste products, respectively. Differences in reference LCIs used, e.g., grid electricity resulted in only slightly different GWP and TPE impact results between the Brodt et al. (
2013) study and the current study.
Overall, the main contributions to the impacts at the facility processing phase observed in the current study (i.e., energy and fuel use) are similar to those observed by Karakaya and Özilgen (
2011) and De Marco et al. (
2018). Across the supply chain, facility processing is the main contributor to total primary energy use—in the current study and in previous studies (e.g., De Marco et al.
2018). One aspect of the processing facility phase which is not assessed in this study is vertical integration with the cultivation phase (and greenhouse phase). However, this aspect of the processing tomato supply chain is worth exploring further given it has proven to increase supply chain energy and water use efficiencies (Benmehaia et al.
2017) and many of the Californian processing facilities are vertically integrated.
Consumer packaging options for both puree and chopped tomatoes include 330–700 g glass bottles with metal or plastic lids, 200 g carton-based containers, and 500–1000 g carton-based containers with plastic caps; and chopped tomatoes containerized in 400–800 g open-top tin-plated steel can with lid (Del Borghi et al.
2014). Given these packaging types and varying sizes for puree and chopped tomatoes, associated fossil GHG emissions range from 0.0743 kg CO
2e per kg tomato puree in a 1000-g carton-based container with a plastic cap to 0.8007 kg CO
2e per kg of chopped tomato in open-top tin-plated steel can with lid (Del Borghi et al.
2014). Based on the Del Borghi et al. (
2014) study, carton-based containers with and without plastic caps have lower impacts than the open top tin and glass bottle. Assuming puree is similar to paste by mass, adding consumer-ready carton-based container packaging impacts to the Californian tomato paste product assessed in this study would increase the global warming potential from fossil-based GHG impacts by 9–11%. However, the total impact from packaging is likely less for the processing facilities in this study, due to packaging their products in large bulk containers (aseptic bags in large re-usable drums or wooden bins) for shipment to food manufacturers. For example, in an earlier study (Brodt et al.
2013), we found that bulk packaging increased the total GHG impacts from transplant production through packaging by less than 2% in the paste product and just under 4% in the diced product.