Data, and sample sources thereof, on water quality life cycle impact assessments pertaining to catchment scale acidification and eutrophication potentials and the benefits of on-farm mitigation strategies

Based on recent spatially aggregated June Agriculture Survey data and site-specific environmental data, information from common farm types in the East of England was sourced and collated. These data were subsequently used as key inputs to a mechanistic environmental modelling tool, the Catchment Systems Model, which predicts environmental damage arising from various farm types and their management strategies. The Catchment Systems Model, which utilises real-world agricultural productivity data (samples and appropriate consent provided within the Mendeley Data repository) is designed to assess not only losses to nature such as nitrate, phosphate, sediment and ammonia, but also to predict how on-farm intervention strategies may affect environmental performance. The data reported within this article provides readers with a detailed inventory of inputs such as fertiliser, outputs including nutrient losses, and impacts to nature for 1782 different scenarios which cover both arable and livestock farming systems. These 1782 scenarios include baseline (i.e., no interventions), business-as-usual (i.e., interventions already implemented in the study area) and optimised (i.e., best-case scenarios) data. Further, using the life cycle assessment (LCA) methodology, the dataset reports acidification and eutrophication potentials for each scenario under two (eutrophication) and three (acidification) impact assessments to offer an insight into the importance of impact assessment choice. Finally, the dataset also provides its readers with percentage changes from baseline to best-case scenario for each farm type.

timised (i.e., best-case scenarios) data. Further, using the life cycle assessment (LCA) methodology, the dataset reports acidification and eutrophication potentials for each scenario under two (eutrophication) and three (acidification) impact assessments to offer an insight into the importance of impact assessment choice. Finally, the dataset also provides its readers with percentage changes from baseline to best-case scenario for each farm type.

Value of the Data
• Life cycle assessment is typically carried out without spatial visualisation of impact assessments · This dataset provides water quality impacts of over 1700 farms (both arable and livestock) which vary from no mitigation strategies to multiple mitigation strategies and demonstrates the benefits of on-farm interventions. The dataset, whilst demonstrating a novel method rather than being a full case-study, could be used by farmers and policymakers to identify where various farming interventions could be deployed geographically to maximise the best management of farming in terms of damage to nature · These data provide a first step to build upon through out-scaling of spatial life cycle assessment to cover water catchments and their potential for water quality improvements, thus enabling stakeholders to target areas where improvements in water quality, or indeed reductions in other environmental impacts, should be prioritised.

Data Description
The dataset comprises 1782 unique farms differentiated by location, intervention (or lack thereof), losses to nature associated with said interventions, and finally their impact assessments. Whilst not all data are used in the main paper, they are reported herein for transparency. For example, soil types 0, 1, and 2 refer to free draining, drained for arable and drained for arable and grass, respectively. Numbers 2 and 3 under rainfall, on the other hand, refer to 60 0-70 0 mm and 70 0-90 0 mm of rainfall per annum, respectively. Each farm ID refers to (a) its location (geographical identification anonymised for farmer protection) and (b) whether it is a predominately arable (1-24) or predominately livestock (25-39) farm. Despite this anonymisation, we have provided end-users with five completed sample surveys for additional transparency, in addition to a consent form which we received from the five participating farmers. Please note that these survey samples are not taken directly from the national June Agriculture Survey (JAS) [3] and are, instead, in-house designed templates to sense-check the validity of licensed (i.e., we do not have permission to share) government survey data. The area for each farm system type calculated under the CSM framework is reported in Column E of the LCI/LCIA dataset, which acts as the functional unit for scaling LCA impact assessments (i.e., acidification potential and eutrophication potentials reported under multiple impact methods). Columns F-K predict the amount of losses to the environment for a range of pollutants for each farm type, whilst Column L reports the carbon stock as tonnes/ha. Energy use (as diesel) is reported in Column M, whilst columns N-P provide the amount of fertiliser used on each farm. Columns Q-Z present the impact assessment results for each farm with Row 1 providing detailed information on the methods used to derive said impacts. Each survey file (i.e., 'Farm 1-5') provides an insight, although incomplete due to licensing reasons outlined above, to demonstrate how data were cross-checked prior to modelling integration within the CSM framework, which ultimately provided the LCI data for the spatial LCIA.

Experimental Design, Materials and Methods
The dataset provided on Mendeley Data was designed using a combination of process-based models (the Catchment Systems Model), deterministic modelling (life cycle assessment) and spatial analysis (Geographical Information Systems). The hypothesis of the study was to test the efficacy of reporting high-resolution (i.e., catchment scale) agri-environmental life cycle assessments in a spatially relevant manner. Data were first collected through a detailed survey of farmers (again, conducted by Hollis et al. [3] , with sample cross-checks provided as complementary assistance to understanding current farm management) to quantify their inputs and outputs and then, subsequently, thousands of scenarios were calculated using the Catchment Systems Model to calculate losses of ammonia, nitrate, nitrous oxide and phosphorus [9][10][11][12][13] . These calculations were subsequently used as a life cycle inventory to calculate three different impact assessments (ReCiPe, Centre for Environmental Management, known as CML, and Environmental Product declarations, known as EPD). Vertically (i.e., by row), the dataset compares various on-farm interventions versus no interventions (baseline) and business-as-usual (BAU; how farmers in the area currently carry out their activities in the real world). Horizontally (i.e., the right-most columns), the dataset compares how choice of life cycle impact assessment affects the results of the different scenarios.

Ethics Statement
Hereby, we (Graham A. McAuliffe, Yusheng Zhang, and Adrian L. Collins) have declared that no human subjects and/or animals were used for the reported research. All authors have no known competing financial interests or personal relationships that could have influenced the work reported in this paper. For the limited bespoke farm surveys, consents were obtained from individual respondents for the use of their returns. All returned farm and personal data were held in strict compliance with UK GDPR regulations and care was taken to ensure that individual farm confidentiality and privacy were maintained throughout data acquisition, storage and use.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data Availability
Data on water quality life cycle impact assessments pertaining to catchment scale acidification and eutrophication potentials and the benefits of on-farm mitigation strategies