Irrigation, harvesting, and P- fertilization contributed the most to the life-cycle environmental impacts of guar agriculture (Fig. 3). Irrigation may offer the most room for improvement of the life-cycle environmental impacts. Research shows that efficient irrigation can also improve yields (Alexander 1988).
The irrigation process has the highest impact in 8 of the 10 impact categories (between 27% and 53%) excluding photochemical ozone formation and respiratory effects. Irrigation impacts are highest in the human health carcinogenic, human health non-carcinogenic, and ecotoxicity categories. Chromium VI emissions to water (74%), zinc emissions to soil (44%) and copper ion emissions to water (45%) are leading contributors to these impact categories. Based on the ecoinvent documentation (Nemecek and Kägi 2007) these impacts may be driven by the upstream mineral extraction and resulting runoff impacts, however no additional information on impact sources of zinc, chromium, and copper was given.
P-fertilization contributed the greatest impacts in eutrophication and respiratory effects with 28% and 25% of the total impacts in those TRACI categories, respectively. Phosphate emissions to surface water (45%) and ground water (39%) and particulates emissions to air (52%) are leading contributors to these impact categories. No data existed for runoff or field emissions; so, it is important to note that all of these impacts result from manufacture and transport of fertilizers. There are estimates of run off and field emissions available in the ecoinvent database for more common European plants (i.e. cotton, rapeseed, wheat, & maize), which are calculated using emissions models SALCA-P (Prasuhn 2006) and SALCA-nitrate (Richner, Oberholzer et al. 2006), but none are available specifically for Guar.
Harvesting also contributes significant impacts across categories, showing the highest impact in the photochemical ozone formation category, contributing 41% of those total impacts. Methane emissions to air is the leading contributor to this impact category, accounting for 96% of the smog impacts. One method for improving the fidelity of this portion of the LCA model is to collect on farm harvest data, detailing specific equipment like the modification used in the combine harvester, in order to get more accurate impacts of harvest.
Interestingly, nitrogen fertilizer had one of the lowest impacts contributing to only about 11% of the total impacts in all the TRACI categories combined. This is a deviation from previous literature of guar agriculture results that show the nitrogen fertilization process as one of the higher impact processes (Gresta, De Luca et al. 2014). The system boundary of the model may have contributed to this difference since no field emissions were found in literature and thus could not be included in the analysis. In addition, specifics of the type and actual percentage of nitrogen being added during cultivation in literature were scarce as discussed in the methods. It is critical that future work understand the efficacies of N fertilizer usage and potential for field emissions.
Another limitation of the existing published data is that it is all from studies involving relatively small plots of land, ranging from 1-3hectares. This resulted in all the results being relevant for small plots but perhaps when scaling up to larger plots the results may not be consistent. One way to improve the results in future studies is to use field data from trial plots and commercial fields to have access to optimized results as well as much more likely commercial farm setting results.
3.3 Sensitivity analysis and scenario analysis
In the sensitivity analysis all of the inputs were varied to evaluate their effect on the overall TRACI impacts, as described in the methods. Irrigation was found to be the largest contributor to environmental impacts overall and it was also the input that the environmental and human health impacts were most sensitive to changing. Changes made to all the other model inputs altered the impacts much less or not at all. A few exceptions include P-fertilization which showed eutrophication, acidification, and respiratory effects to be most sensitive to its variation. N-fertilization, despite contributing so little to overall impacts, resulted in resource depletion to be as sensitive to its variation as irrigation.
Two scenario analyses were carried out; 1) irrigation: where minimum and maximum literature values were compared to the baseline average of literature values, and 2) nitrogen fertilizer: where alternative fertilizer types were evaluated. The irrigation scenario compared minimum and maximum irrigation literature values to the baseline. As expected, the results show that less water results in less impacts (Fig. 4). In this scenario, impacts were especially reduced in ecotoxicity, GWP, and resource depletion when the minimum irrigation value was used.
One thing to note is that this study only uses sprinkler irrigation in the analyses. This is a result of the available literature data only including sprinkler irrigation values. It is possible that other irrigation methods, like drip or flood, could have varying impacts. One comparison study by Eranki et al. for another desert crop, guayule, shows that drip irrigation was much more efficient in terms of water applied and yield then flood irrigation, and it also used less energy consumption and produced less environmental impacts (Eranki 2017). Perhaps similar field trial studies could provide data on which irrigation method is most efficient for guar cultivation. Field trial studies could also potentially provide enough data to support the development of a guar specific irrigation model instead of using the generic ecoinvent sprinkler irrigation impacts, which could provide guar specific impacts for each irrigation method investigated.
The nitrogen fertilizer scenario was conducted because of the great lack of detail provided in literature on N-fertilizer values. Many of the sources did not provide fertilizer types, brands, compositions, or N percentage. Urea was used as the baseline fertilizer in this study because it is incredibly common in agriculture. The scenario analysis compared urea to three other common fertilizers: monoammonium phosphate, calcium ammonium nitrate, and ammonium nitrate. The results in Fig. 5 show that across all the scenarios the life cycle impact categories that have the largest variation are acidification, ecotoxicity, eutrophication, and global warming impacts. When compared to the baseline N-fertilizer Urea, the only scenario that has a lower total impact is monoammonium phosphate. Using monoammonium phosphate decreases the impact in every impact category except acidification (increasing by < 1 %) and eutrophication (increasing by 5%). This decrease in total impacts can be contributed greatly to the composition of monoammonium phosphate. It contains both nitrogen and phosphorus and therefore using it for N fertilizer can also offset some of the need for adding P fertilizer. The other two scenarios (calcium ammonium nitrate & ammonium nitrate) increased the impacts in every impact category except for a 1% decrease in eutrophication when using ammonium nitrate. Ultimately this scenario analysis shows that using multinutrient fertilizers like monoammonium phosphate, that have both N and P within its composition may be the most efficient way to fertilize guar. Though these are promising preliminary results, it is important to measure field emissions and incorporate them into the analysis for future studies, which may significantly impact the overall results.