Introduction

Mitigating greenhouse gas (GHG) emissions has become an urgent global imperative to curb the rise in temperatures and limit the adverse effects of climate change. The Intergovernmental Panel on Climate Change (IPCC) has underscored the necessity of substantial reductions in GHG emissions to restrict the global temperature increase to 1.5 °C by 2050 and maintain it well below 2 °C above pre-industrial levels by 2100 (IPCC 2018, 2023). Globally, agriculture is responsible for more than 50% of the non-carbon dioxide (CO2) GHG emissions (Jia et al. 2019), mainly from biogenic GHG emissions (BGE) of methane (CH4) and nitrous oxide (N2O).

In contrast to most developed countries, Aotearoa-New Zealand’s (New Zealand) GHG emissions are dominated by agriculture, mainly from livestock farming, contributing approximately 53% of gross emissions (Ministry for the Environment 2024). In 2021, nearly 90% of agricultural emissions originated from ruminant livestock. To help reduce these large livestock BGE, New Zealand has set statutory 2050 targets for biogenic CH4 emission reductions of 24–47% from 2017 levels, the final target depending on a range of factorsFootnote 1. There are no targets for biogenic N2O. However, reducing agricultural GHG emissions may come at a cost to the New Zealand economy. Food and fibre production accounts for about 80% of total export revenue, 15% of employment and 11% of gross domestic product (Ministry for Primary Industries 2019, 2023), at the same time supporting the social, cultural and economic vitality of rural communities (Cradock-Henry et al. 2020).

Overall, livestock farming accounts for about 60% of export revenue in New Zealand, while horticulture and forestry account for about 15% each. Pastures make up about one-third (9.4 M ha) of New Zealand’s total land area, production forestry is about 1.7 M ha, while horticultural and arable crops account for approximately 0.5 M ha (Stats NZ 2021a). Export revenue from food and fibre has been increasing. Between 2017 and 2022, export revenue increased from NZ$38 billion to NZ$53 billion, with the New Zealand Government aiming to capture even more of food and fibre economic potential with various socio-economic and sustainability targets for 2030 (Ministry for Primary Industries 2020).

New Zealand was to be the first country to price agricultural emissions from 2025 to incentivise farmers to reduce their BGE (Ministry for the Environment 2022); this is postponed until at least 2030 following a change in the Government in 2023. Most countries have resisted applying stringent policies to reduce BGE from agriculture (Leahy et al. 2020), relying on non-policy mechanisms instead (Renwick and Wreford 2011). The New Zealand Government’s first emissions reduction plan (Ministry for the Environment 2022) set out new initiatives to reduce agricultural BGE and enhance CO2 uptake by plantation and native forest. The plan includes pricing agricultural emissions, accelerating mitigation technologies, supporting producers to make changes and transitioning to lower-emissions land uses and systems.

Despite reductions in per animal BGE intensity (kg CO2-equivalent/kg product), largely through improved animal performance, BGE have remained static since 2008, and compared to 1990 have increased (Leahy et al. 2019). There is an expectation that new mitigation technologies and their uptake, together with changes to livestock farm systems, will achieve large CH4 reductions (Ministry for the Environment 2022). While promising technologies are being explored, there will be challenges to make them successful at a farm scale (Reisinger et al. 2018), and there are also concerns that estimations of the efficacy of these technologies are overly optimistic (Leahy et al. 2020). Hence, land use change will be an important component of the mitigation portfolio for reducing the impact of climate change (Committee on Climate Change 2018). Historically, forestry to sequester carbon has been important for reducing New Zealand’s net GHG emissions. However, relying on afforestation alone to meet future GHG emissions target ambitions is risky and reduction in BGE is required (Climate Change Commission 2023; Reisinger et al. 2018).

Changes from livestock farming to non-forestry land uses, such as arable and horticulture, have received less attention as a mitigation tool, despite economic returns for some of these land uses being greater. While the review by Reisinger et al. (2018) of New Zealand agriculture excluded any assessment of land use change to cropping systems in their mitigation options, two recent modelling studies exploring different land use scenarios concluded that transitioning from livestock to forestry and cropping would reduce emissions to help meet 2030 and 2050 climate emission targets. However, both studies identified that government policy interventions were needed to encourage land use change (Climate Change Commission 2021; Dorner et al. 2018).

This study contributes to the relatively scarce literature on the mitigation potential of changing to various cropping enterprises to reduce agricultural GHG emissions. We employ a novel approach that integrates market opportunities, economic returns and GHG emissions. Unlike other approaches which typically begin with land suitability assessments, we start by identifying promising crops and products based on demand and market growth opportunities for high-value exports (Coriolis 2015). Our objectives were to assess: whether changing from livestock farming to high-value crops is a realistic GHG mitigation option for New Zealand farming; what is the potential size of reduction of emissions; and what, if any, economic benefits result from change. Furthermore, would there be greater benefits from targeting change to specific regions? Using exemplar land use change scenarios, we compared the outcomes of focusing on a single region versus distributing similar changes across New Zealand. Our results are intended to provide quantitative insights to enrich discussions about the potential impacts of land use changes on BGE and economic returns.

Method

Our approach was designed around a conceptual framework shown in Fig. 1. From an analysis of export market opportunities, we identified promising high-value crops or products we named “alternative crops” (Identifying alternative crops and products). Then, we predicted where these alternative crops were suited to grow and estimated their GHG emissions and profitability. We used the Land-Use Management Support System (LUMASS) spatial optimisation model (Herzig et al. 2013) to allocate alternative crops onto pastoral land. Outputs from LUMASS were the new GHG emissions, changes in GHG emissions, new profitability and changes in profitability. We tested the framework to explore potential changes in national and regional GHG emissions against New Zealand’s 2017 agricultural GHG emissions and 2050 CH4 emission reduction targets, and assessed potential economic benefits. Economic values for 2017, the base year for New Zealand’s BGE reduction targets, were used to estimate profitability.

Fig. 1
figure 1

Framework used for estimating the effects of changing land uses on on-farm or orchard greenhouse gas emission (GHG) and profitability ($). GIS, Geographic Information System. LUMASS - spatial modelling and optimisation. Activity data refers to input data used to calculate GHG emissions (e.g. fertiliser and residue nitrogen). Emission factors (EFs) are applied to the activity data to estimate the fraction of nitrous oxide or methane emitted following the IPCC methodology used for national greenhouse gas inventories

Full size image

Identifying alternative crops and products

Data analysis focussed on current market patterns, demand trends and international trade. In advance, we set a target of 12 alternative crops. This was considered a manageable number for our analysis and sufficient to capture the best alternative crop options. The steps were:

  1. 1.

    Identifying a long list of potential crop and product options from a review of literature, including papers and reports from New Zealand and overseas research organisations, industry and government. The review considered volumes and values of production and consumption (export and domestic) and qualitative aspects such as discussion around the key trends, degree of coverage and attention given to products. From the review, we identified key drivers of demand in the global fruit and vegetable markets.

  2. 2.

    Assessing the export markets best placed to represent demand for New Zealand products from a literature review of global trends and trade access. Data sources included New Zealand Ministry of Foreign Affairs and Trade research, New Zealand trade agreements (current and pending) and current New Zealand Government trade statistics.

  3. 3.

    Analysing United Nations ComTrade import and export trade data for 2012–2016 (https://comtrade.un.org/data/) to assess product demand and growth, the key international markets and verify our analysis in previous steps.

  4. 4.

    Confirming the 12 alternative crops with Ministry for Primary Industry (MPI) staff before progressing further.

From the initial review of literature, six key selection criteria were identified and used to select 32 potential crops or products: (1) growth opportunity in Asia; (2) premium or high-value product; (3) New Zealand export growth or volume; (4) popular in processed food; 5) New Zealand brand recognition or position and (6) unique perceived health benefits. Analysis of existing trade partnerships, market importance, trade agreements and ComTrade data led to selecting 14 target or favourable markets from 34 starting possibilities. Twelve target alternative crops were identified from the literature reviews (Fig. 2) and validated by the trade data analysis. Analysis of ComTrade data identified several other promising crops and products worth further investigation (Supplemental Material Table S1). Ten of the 12 crops and products had been identified in previous studies as either export leaders or emerging opportunities (Coriolis 2014, 2017).

Fig. 2
figure 2

Promising high-value crops to be used in land use change modelling identified through market analysis. Key below each crop name indicates different assessment criteria used

Full size image

Growing suitability

Growing suitability for a specific crop was estimated from key climate (temperature and rainfall regimes), soil (soil depth, drainage and pH) and terrain (slope and aspect) attributes. Crop-specific suitability attributes were identified and crop responses to these attributes defined, based on literature and expert knowledge (Kidd et al. 2015; Vetharaniam et al. 2022a). The number and types of rules varied with each crop. For example, some temperature rules were based on crop phenological periods when plants are sensitive to frost or require chilling. Two complementary GIS approaches were used to create crop suitability maps and data tables (Fig. 1) and are described in the Supplementary Material. Examples of the outputs for growing conditions and land use suitability are given in Supplemental Material Table S2 and Fig. S1Footnote 2.

Farm and orchard greenhouse gas emissions

We limited the GHG emission estimates to those occurring on-farm or orchard, or indirectly from volatilised, leached or run-off nitrogen emitted from the farm or orchard, i.e. we did not include “upstream” emissions related to the manufacture of fertiliser or other agrichemicals. Although our focus was primarily on farm BGE, we also estimated the CO2 emitted from urea, lime and dolomite applications to the soil. This was particularly important for truffles, which require high soil pH conditions often produced through large lime additions. Emissions from cropping and livestock (dairy, sheep and beef) were calculated following New Zealand’s GHG inventory methodology (Ministry for Primary Industries 2022a; Thomas et al. 2011). Direct and indirect N2O emissions from crops were estimated from nitrogen (N) inputs from synthetic fertilisers, returns of crop residues and livestock excreta deposition and off-field manure management. Estimates of CH4 emissions from ruminants and manure management were based on the New Zealand inventory data for each livestock type based on feed intake. Most of the alternative crops are not included in the New Zealand Inventory (Ministry for Primary Industries 2022a). Currently, only 14 “key crops” are reported (Ministry for the Environment 2022) and none of these is tree or perennial fruit crops. Therefore, it was necessary to estimate the amounts of residues returned for crops not included in the Inventory, as well as the amounts of synthetic fertiliser, lime and dolomite applied for all the crops. This information was collected from industry experts, production statistics and literature review. To enable comparison of GHG emissions between land uses, N2O and CH4 were converted to CO2 equivalents (CO2-e) using global warming potentials of 298 and 25, respectively, over a 100-year horizon (GWP100) (IPCC 2007), the values used in the New Zealand inventory before 2024. GHG emissions are reported on a CO2-e per ha basis. For livestock, BGE were first calculated on a per animal basis and then converted to a per ha basis using average regional animal stocking rate data (Stats NZ 2017).

Estimation of profitability

Potential profitability (NZ$ per ha) was estimated for each of the alternative crops and livestock enterprises. Financial data were sourced from a combination of government and industry statistics, bank publications, other sources such as budget handbooks and internet news items and discussion with industry experts. For crops not widely grown in New Zealand, estimates of profitability were based on limited information. Consequently, the level of confidence placed in the estimates varied across the crops. Given the disparate nature of the sources of data, a major challenge was to ensure that the measures of profitability were comparable across the alternative crops as well as with the existing livestock enterprises. For example, some were presented as gross margins, while other included fuller costings such as cash operating surplus or earnings before interest and tax. This challenge was exacerbated by the fact that the types of crops considered varied significantly (e.g. annual versus perennial). This meant that there were significant differences in the levels of establishment costs, time to maturity and productive life. While recognising the importance of these factors in determining the overall returns on investment, for the purposes of comparison with existing livestock enterprises, it was decided to use the annual costs and returns when mature production levels are reached as a measure of profitability. Since the available financial data were largely based on enterprises growing in well-suited environments, we did not attempt to allocate different profitability values to the different land suitability classes. We estimated establishment costs and time to maturity based on available data, to provide further information concerning the relative performance of the crops, but this was not included in estimates of profitability (Supplemental Material Table S4). As for GHG emissions, our baseline year for profitability estimates was 2017.

Scenario development

We set up our scenarios in consultation with MPI staff responsible for informing land use and agricultural emission policy, to see what impact land use change of our selected crops would have on the 2050 reduction target of 24–47% of New Zealand’s CH4 emissions in 2017. Biogenic CH4 emissions were 26,210 kt CO2-e of the total agricultural BGE (37,119 kt CO2-e) in 2017 (Ministry for the Environment 2019); therefore, 24% and 47% reductions of biogenic CH4 are equivalent to approximately 7000 and 13,700 kt CO2-e, respectively. For projecting growth of potential land uses, we reviewed available sector growth strategies, examined previous studies, interviewed sector experts and evaluated past growth from New Zealand’s agricultural production statistics (Stats NZ 2017). The 2017 baseline land use information was primarily sourced from the Stats NZ (2017) agricultural production census, which provides data at regional levels (Stats NZ 2017). We primarily focussed on land use patterns between 1997 and 2017. In the absence of industry growth publications, we contacted industry representatives. Future land use projections for 2050 used in all scenarios are described in Table 1. For most crops, we assumed a doubling of land in 2050.

Table 1 Future land area growth scenarios for the different land uses for 2050
Full size table

Justifying projected growth in areas of the enterprises to 2050 was challenging, especially for those land uses not yet established in New Zealand. In our initial analysis, it was clear that even those crop sectors with a growth strategy, e.g. apples and growth in output, were expected through intensification (i.e. higher yields per ha) as well as through extensive area expansion (i.e. more cultivated ha), making it hard to separate out the likely impacts of the two. Between 2002 and 2017, the horticultural land area grew by about 1.3% per year (Stats NZ 2021a). Some crops have grown much faster, such as wine grapes for which the area grew at a rate of 22% per year between 2000 and 2010 (an average of 2350 ha/year). Since 2017, the wine grape area has grown at 2.1% per year (1100 ha/year) (Plant and Food Research 2021). Avocado and cherry growing areas have also increased markedly. Between 1997 and 2017 the area in cherries grew by 3.7% per year, while avocados increased nearly fivefold between 1999 and 2011, although since 2011 the area in avocado is now to about four times its size in 1999 (Plant and Food Research 2021). In contrast, for other sectors, such as kiwifruit and potato, growing areas have remained relatively static over the same period (Plant and Food Research 2021). Given the opportunities for the crops that we considered, it is realistic that the areas could double by 2050. This implies a growth rate of around 4.7% per year. Since there were no available data for truffles, we assumed that the industry becomes established during this period and could reach 2000 ha by 2050. We also assumed that the nut industry will expand to reach a reasonable scale of 4000 ha. We assumed an increase of 100,000 ha of mānuka plantation from the current 16,800 ha, less than industry estimates of 184,000 ha by 2035 (Mānuka Research Partnership 2019). Hence, almost half of the change in land use in our scenarios will be in mānuka trees. If these 2050 projections were achieved, it would increase the alternative crops by c. 0.195 M ha. To put this in context, arable and horticultural crops occupied about 0.476 M ha in 2017. This compares to 6.864 M ha of high production pasture and c. 1.74 M ha of mainly Pinus radiata plantation forestry (Ministry for Primary Industries 2022b).

Optimising land use allocation

We used LUMASSFootnote 3 (Herzig et al. 2013; Herzig et al. 2018) and a land use map (Manderson et al. 2018) resampled to a 500 × 500-m grid to explore three scenarios to assess how replacing livestock farming with the alternative crops affected GHG emissions and profitability. Data inputs to the model were (i) the alternative crop suitability geospatial map layers and associated tables of emissions (total GHG and CH4 only) and profitability, and (ii) geospatial information on the current location of pastoral land uses plus their GHG emissions and profitability. LUMASS enabled us to (i) allocate alternative crops based on their suitability criteria; (ii) define multiple optimisation objectives, e.g. land use suitability and GHG emissions; (iii) specify and rank multiple objectives; (iv) ensure that the increase in alternative crop land would occur only on land currently used for livestock, i.e. not replacing existing high-value crops and (v) generate maps and tables of aggregated emissions and profitability.

Scenario 1 “National” was designed to estimate the potential reduction in agricultural GHG emissions and profitability impacts by allocating alternative crops to current pastoral land based on the land most suited for the crops. Scenario 1 optimised land use changes for each region and then GHG emissions and profitability were aggregated for the whole country. In contrast, scenarios 2 and 3 were designed to provide greater insight into the GHG emission reductions and economic impact from land use change at a regional scale. The Canterbury region of the South Island was chosen because it has the highest dairying stocking rates and BGE/ha and the largest area of crops in New Zealand (Stats NZ 2021a). It has also undergone large-scale land use change to intensive irrigated dairy since the 1990s (Pangborn et al. 2015). Between 2002 and 2017, dairying increased by approximately 245,000 ha (Stats NZ 2021a). Scenarios 2 and 3 differed in the way that land use change was optimised. In Scenario 2 “Canterbury Unconstrained”, change was optimised to allocate alternative crops to the best pastoral land. Effectively this meant that most change to alternative crops occurred on dairy pasture, which tended to be land suitable for a wide range of productive uses. In contrast, in Scenario 3 “Canterbury Constrained” the optimisation was restricted to include no more than 50% dairy pasture, effectively forcing more land allocation to occur on sheep and beef pasture.

Results

Based on crop growing requirements, availability of suitable land is unlikely to constrain transitioning to alternative crops, although regional climates will limit the range of crop options. For example, avocados were limited to the upper North Island due to frost risk. Kiwifruit was unsuited in the lower South Island, although some varieties are suited to warmer South Island regions where they are currently not commercially grown. In contrast, apple growing is widely suited to many regions in New Zealand due to the wide range of available cultivars (Vetharaniam et al. 2022a). Areas well suited for growing the alternative crops are summarised by region in Supplemental Material Table S5.

Greenhouse gas emissions

GHG emissions from the alternative crops were driven by the N inputs from fertiliser and crop residues, with the largest from potatoes (1900 kg CO2-e/ha) which had the largest fertiliser N inputs (250 kg N/ha) (Table 2 and Supplemental Material Table S6). GHG emissions from kiwifruit and onions were the next largest (c. 1250 kg CO2-e/ha). Mānuka honey and chestnuts were the lowest (0–100 kg CO2-e/ha) where N inputs from fertiliser (0–20 kg N/ha) or residue inputs were nil or small. GHG emissions from truffles were largely due to CO2 emitted following lime applications to maintain favourable pH for truffle growth. Overall, GHG emissions from alternative crops were much lower than from livestock (Table 2), where the greatest contribution (> 75%) was from CH4. Based on the national average stocking rate for dairy of 2.78 cows/ha, GHG emissions were 8198 kg CO2-e/ha for dairy, while for beef cattle they were 2614 kg CO2-e/ha (1.41 cattle/ha) and 2188 kg CO2-e/ha for sheep (6.27 sheep/ha). Stocking rates varied regionally. The highest stocking rates were in Canterbury with an average of 3.64 cows/ha and an average emission of 10,743 kg CO2-e/ha.

Table 2 Greenhouse gas emissions from alternative high-value crops and livestock. Nitrous oxide and methane were converted to CO2-equivalents (CO2-e) using global warming potentials of 298 and 25, respectively, over a 100-year horizon (GWP100) (IPCC 2007)
Full size table

Estimation of profitability

Profitability estimates for the alternative crops and livestock are provided in Table 3 and Supplementary Material Table S7. Except for mānuka honey and peas, alternative crops were expected to be more profitable per ha than livestock. Kiwifruit (Zespri™ SunGold™) and blueberries were predicted to be the most profitable (> NZ$60,000/ha), followed by cherries, truffles and avocado (NZ$27,000 to NZ$40,000/ha). Profitability estimates for winegrapes and apples were lower (c. NZ$15,000/ha) but still several times greater than profit estimates for dairy (NZ$1,800/ha) and for sheep and beef (NZ$337/ha). Profitability of potatoes and onions crops was similar (NZ$5385 to NZ$6331/ha). Sources of financial information for several of the land uses were sparse. We considered the estimates least robust for cherries, blueberries, truffles and chestnuts, whereas we were more confident in the data for the large, established export enterprises of dairy, sheep and beef, wine, apples and kiwifruit.

Table 3 Assumed profitability for each alternative crop or land use (NZ$/ha) based on 2017 data. To provide an indicator of the confidence that can be placed in the figures, the final column (Robustness) uses a star system where four stars indicate that the figures are based on official statistics and one star indicates that data have come from informal sources (websites, news items, etc.)
Full size table

Land use change scenarios

Scenario 1 “National” resulted in a relatively small reduction in GHG emissions of 370 kt CO2-e, of which 349 kt CO2-e was CH4 (Table 4). This is equivalent to a 1% reduction in total GHG emissions and 1.2% of CH4 emissions from New Zealand agriculture in 2017 (Fig. 3, Table 4). As a contribution to the New Zealand biogenic CH4 emission reduction 2050 targets, this is between 2.5% and 5.0%, depending on the lower or upper target value. In contrast, for Scenarios 2 and 3, GHG emission reductions were much larger when all the land use change occurred in Canterbury. Changing approximately 195,000 ha of pastoral livestock land to alternative crops resulted in a 28 to 33% reduction in Canterbury agricultural GHG emissions, and importantly, this also resulted in about a fivefold greater reduction in New Zealand agricultural GHG emissions compared to Scenario 1 “National” (Table 4). Given that the total area of land use change was the same for all the scenarios, this difference largely reflects the effect of reducing relatively large areas of intensive dairying in Canterbury. The greatest reduction was in the Scenario 2 “Canterbury Unconstrained”, in which 90% of land use change was simulated on dairy land (175,330 ha). This is equivalent to almost halving the area of dairying in Canterbury (359,081 ha in 2017; Stats NZ 2021a), compared to a reduction of about 97,500 ha of dairy land in Scenario 3 ”Canterbury Constrained”. Fig. 4 shows the allocation of alternative land uses on the livestock area in Canterbury based on allocation to maximise the reduction in BGE (Scenario 2 “Canterbury Unconstrained”). Such change in this scenario would contribute 12% to 23% of the New Zealand CH4 emission reduction targets (Table 4).

Table 4 Estimates of reductions in national agricultural greenhouse gas emissions (GHG), livestock methane (CH4) emissions (kt CO2-e and % of total) and increases in profitability (NZ$ million and % of total) for New Zealand and Canterbury (results in parentheses) for three land use change scenarios. LUMASS spatial optimisation model was used to allocate c. 195,000 ha of alternative high-value crop land uses to existing livestock land
Full size table
Fig. 3
figure 3

Changes in greenhouse gas (GHG) emissions and profitability ($NZ million) estimated from an optimistic scenario of doubling the area of high-value “alternative crops” across New Zealand and allocating them to land that is farmed with livestock (scenario 1). The most suitable land was allocated to grow the alternative crops. Note that GHG emissions from mānuka are low due to the low nitrogen inputs from residues used for calculating emissions; it was assumed that no fertiliser was applied

Full size image
Fig. 4
figure 4

Simulated land use change to approximately 195,000 ha of alternative high-value, low greenhouse gas emitting crops, replacing dairy, sheep and beef land in the Canterbury region. The Land-Use Management Support System (LUMASS) spatial optimisation model was constrained to allocate no more than 50% of dairy land to the alternative crops (scenario 3, Table 1). The insert shows the location of Canterbury region in New Zealand

Full size image

Greater economic benefits were predicted from changing to the alternative crops in all the scenarios, compared with maintaining current land uses. Profitability was predicted to increase by NZ$1.32 billion (Fig. 3), an increase of about NZ$6800/ha compared to average returns for livestock that were between about NZ$340 and NZ$1800/ha (Table 3). Although not directly comparable, the increase in profitability is equivalent in value to about 4% of New Zealand’s agricultural export earnings in 2017. Canterbury Scenarios 2 and 3, which provided the greatest reduction in BGE, were less profitable than Scenario 1 “National” since most of the changes were allocated to land used for dairying (Table 4). Nevertheless, based on the assumptions used to estimate profitability, this would still make the land 73% and 81% more profitable in a crop enterprise than in its current use in Canterbury Scenarios 2 and 3, respectively. The area of land allocated to mānuka (100,000 ha) in the scenarios had a relatively large effect on both the overall profitability and GHG emission reductions from the new alternative crop area (195,000 ha) because of low estimates of both profit and GHG emissions, i.e. land converted from dairy would be less profitable but would have the lowest emissions of all the options.

Discussion

The GIS-based framework was developed with policymakers in mind to help inform potential trade-offs of land use change. It allowed us to explore GHG emissions and economic trade-offs from changing land use, and to run scenarios to explore benefits of more regionally specific change. Examining the market demand trends and opportunities helped identify realistic crop options to base our growth scenarios. Evidently, there is the opportunity for significant growth in the alternative crops, and mostly, they would be a more profitable use of the land. Furthermore, changing to lower-emitting, high-value land uses could be an effective tool to reduce BGE and CH4 emissions, reducing the reliance upon other CH4 mitigation technologies. If changes were focussed on areas with the greatest BGE, such as irrigated dairy farming in Canterbury, greater reduction benefits might occur. Comparatively, the rapid conversion of over 245,000 ha of land to dairy farming in Canterbury from 2002 to 2017 is much greater than our projection of 195,000 ha. Our analysis showed that finding suitable land would not be a constraint for the alternative crops. Clothier et al. (2017) estimated that 2.1 M ha was suited to high-value horticultural crops, a magnitude greater than our scenarios. Dorner et al. (2018) modelled a more optimistic expansion of horticultural and arable crops of up to 1 M ha by 2050, doubling the current area of all crops. Our scenarios better align with crop area increases of 1400 to 2000 ha per year to 2050 modelled by the Climate Change Commission (Climate Change Commission 2021, 2023). If 100,000 ha of mānuka is planted by 2050, well within projections of 6000 ha/year between 2019 and 2035 (Mānuka Research Partnership 2019), our optimistic projection of land use change to the other alternative crops of about 3300 ha per year is reasonable.

Similarly, the economic benefits of transitioning to high-value crops are compelling, with tree crops and berries returning profits of more than NZ$8000/ha and some > NZ$60,000/ha. More recent economic data for some of these crops suggest similar, or in some cases greater profitability, with estimates of NZ$65,000/ha and $106,000/ha for blueberries and kiwifruit (Zespri™ SunGold™), respectively (Harris 2024). Increasing and diversifying food exports would also provide greater resilience to export market volatility, an outcome sought by the New Zealand Government (Ministry for Primary Industries 2020). Fluctuations in livestock export prices impact the New Zealand economy, where livestock products account for about 60% of export revenue (Ministry for Primary Industries 2023). While the benefits associated with transitioning to alternative crops appear compelling, there are a range of other considerations.

Future analyses should consider the large, upfront capital investment required to establish some of the fruit and tree crops. We considered how this might be included in our analyses; however, there was insufficient or unreliable financial information to make meaningful comparisons between what are complex and diverse livestock and cropping farming systems. Hence, we assumed that the alternative crops were already established and profitable. Indicative capital investment requirements and time to production for the various crops were explored and are provided in Supplementary Material Table S4. If we had been able to include the upfront costs, the relative economic returns of the different crops would less, reducing their attractiveness to landowners. Nor did we consider volatility in production due to weather or market dynamics that would affect returns year to year. This would be an important consideration given the high capital investment to establish the new crop enterprises. Similarly, we did not attempt to include the effect of New Zealand’s emission trading pricing in our economic analysis, i.e. accounting for the economic benefit of carbon credits from growing “non-food producing” trees such as mānuka and truffle trees or factoring the pricing of BGE to see if this makes land use change to crops more attractive.

Two other factors that could make land use change less attractive, particularly for the less established crops (e.g. chestnuts and truffles), are the lack of effective supply chains and being competitive in new international markets. Major investment would be required to develop the necessary supply chain infrastructure such as processing facilities and transport networks. In contrast, livestock products and established export crops such as kiwifruit, apples and wine have existing well-developed, sophisticated and efficient supply chains.

Future analysis should also consider GHG emissions from the foods’ supply chains. We ignored CO2 emissions from energy use within the farm or orchard boundary or through the supply chains. Previous studies for arable and perennial crops suggest that emissions from fossil fuel combustion are likely to be relatively small, accounting for 10 to 35% of total on-farm GHG emissions and less than the emissions from fertiliser and residues (Barber et al. 2011; Mithraratne et al. 2008). Off-farm GHG emissions may be much more important. However, accounting for these emissions is complex and requires knowledge about processing, transport networks and markets. Currently, this type of information exists only for a small range of export food products (e.g. Majumdar and McLaren 2024; Mazzetto et al. 2021; McLaren et al. 2021). We did not account for CO2 emissions related to changes to soil carbon or plant biomass due to land use change. Simple approaches used in the New Zealand GHG inventory were too coarse and too uncertain, i.e. perennial crops are considered a single land use despite large diversity in vines, shrubs and trees, understories and shelter. Whitehead et al. (2021) noted there is large uncertainty in both the stocks and changes of soil carbon due to land use change and management.

Effects of climate change that result in shifts in crop suitability and distribution (Ausseil et al. 2021; Tait et al. 2018; Vetharaniam et al. 2022b) will become more important in land use decisions. Changes in rainfall distribution and seasonality may lead to new water availability constraints, particularly as most alternative crops require irrigation. Notably the highly intensive dairy farming in Canterbury, the area with highest BGE/ha, is reliant on irrigation and accounts for nearly 60% of the irrigated land in the region (Stats NZ 2021b). Conversion to alternative crops would be easier where there is already a reliable irrigation supply. Moreover, land use change from irrigated pasture at crops could free up water resources for other uses; water requirements for crops tend to be much less than pasture and have greater economic returns per volume of irrigation water.

Availability of skilled labour maybe a barrier to change. High-value crops require more labour than livestock. At least twice as many staff are required on a full-time basis for horticulture (46.8 FTE/1000 ha) compared to dairy (21.3 FTE/1000 ha) (Dorner et al. 2018). As well as requiring more staff for crop production, the work tends to be much more seasonal than livestock farming. Shortage of seasonal labour has been an issue. To address this, since 2007 New Zealand operates a Recognised Seasonal Employer scheme that enables employers within the horticulture and viticulture industries to recruit overseas workers mainly from Pacific countries (Horticulture New Zealand 2022). Lack of labour for horticultural crops was particularly highlighted by COVID travel restrictions.

Recently, the New Zealand Government released “Fit for a Better World”, a 10-year plan aimed to accelerate productivity, sustainability and inclusivity in the primary sector (Ministry for Primary Industries 2020). The plan aims to address some of the identified barriers that are inhibiting land use change (Journeaux et al. 2017) and meanwhile encourage innovation. Actions in the plan include maintaining and building market opportunities, increasing employment opportunities and revitalising rural communities, increasing irrigation water storage, breeding more high-value plant varieties and supporting development of new high growth enterprises. The National Policy Statement for Freshwater Management (Ministry for the Environment 2020) targeted at improving freshwater quality is also likely to drive land use change and reduce livestock farming intensity (Djanibekov and Wiercinski 2020; Ministry for the Environment and Stats NZ 2017).

Ultimately, understanding the human dimension of land use decisions is important. It is highly complex, difficult to predict and requires more investigation (Renwick et al. 2022; Cradock-Henry et al. 2020). For livestock farmers, managing animals may be perceived as less risky than cropping where the risk of crop failure and loss of income may be too large a barrier to change (Jaffe 2017). Other farmers may simply identify with farming animals and have no interest in changing to non-livestock systems (Cradock-Henry et al. 2020).

Conclusions

Transitioning from livestock to high-value crops presents a significant opportunity for mitigating GHG emissions. Even modest changes in land use could make a substantial contribution to New Zealand’s ambitious 2050 CH4 emission reduction targets. Moreover, expanding and diversifying non-livestock export food products could enhance resilience to market volatility. Our framework has provided valuable insights into the trade-offs between GHG emissions and economics at both regional and national scales. For instance, policies that facilitate changes on intensively grazed lands could yield the greatest emission reductions while remaining economically viable compared to current land uses.

Obtaining reliable economic data for growing new high-value crops poses challenges, particularly for less established crops. Selecting alternative crops based on a comprehensive analysis of market demand trends and growth opportunities for New Zealand products provided valuable insights. A deep understanding of opportunities will be crucial for successfully exporting high-value products to existing and new markets. Even with promising export opportunities, establishing alternative crops and developing new supply chains will be challenging. Government initiatives (e.g. Fit for a Better World) to overcome a myriad of complex economic, infrastructural and social barriers, including farmers’ willingness to change, and enable innovation will be important. The availability of suitable land for alternative crops is not a limiting factor.

Future modelling efforts should consider changes in soil carbon and crop biomass, incorporate net GHG emission pricing and include establishment costs associated with transitioning to new land use enterprises. Furthermore, there is an opportunity to expand the framework to encompass other environmental factors such as contaminant losses to water or other relevant indicators.