A Risky and Potentially Costly Future: Implications of Climate-Induced Changes in Groundwater and Flooding for Coastal Dairy Farming in New Zealand

Abstract

Climate change poses significant risks to New Zealand’s coastal agriculture through both slow-onset hazards (e.g., gradual sea level-induced groundwater rise) and sudden-onset hazards (e.g., increasing frequency and severity of storms). These physical changes threaten the productivity and economic viability of coastal farms. However, few studies assess their combined economic impacts in a manner that supports land-use planning. This paper presents a conceptual framework to examine the implications of interacting slow- and sudden-onset climate hazards for New Zealand dairy farms, informed by real-world consultation with subject-matter experts to support assessment. We draw conclusions that illustrate the monetary impacts on farms associated with potential absorptive, adaptive, and transformational responses. The findings highlight the critical role of timing as environmental conditions deteriorate under climate change, as well as the need for policy frameworks that recognise and monetize the contribution of ecosystem services provided by coastal vegetation habitats to social, cultural, and environmental wellbeing. Incorporating these values into present-day financial decision-making is essential for supporting climate-related financial risk reduction and long-term land-use planning. Without such frameworks, the most beneficial land-use transitions are unlikely to be affordable or sustainable in New Zealand, especially towards the year 2100.

1. Introduction

Global climate change poses significant threats from sea level rise and extreme weather events, such as flooding and droughts [1,2]. These phenomena have the potential to damage infrastructure and displace millions of people worldwide [3]. Challenges may be felt especially in coastal lowlands where both sea level rise and extreme oceanic and terrestrial weather events can coincide to intensify impacts [4,5].

Globally, the rate of mean sea level rise is accelerating, increasing from approximately 2.1 mm/year in 1993 to around 4.5 mm/year in 2023 (estimated) [6], and sea-level is estimated to rise by approximately 60 cm to 1 m or more by the year 2100 [7,8,9]. Currently, around 10 per cent of the world’s population are located in low-lying coastal areas and this is expected to rise to one billion by 2050 [7]. The consequences of climate change systematically affect human assets and economic models, impacting livelihoods and economic stability on a global scale [10]. Research indicates that real Gross Domestic Product (GDP) is projected to decline globally as a result of extreme weather events, drought, and rising sea levels [11]. However, the extent of future sea-level rise, the resulting impacts on low-elevation coastal zones and the capacity for global societies to adapt remains uncertain [3]. Additionally, knowledge of the consequences of sea level rise and associated groundwater rise on local economies is a significant research gap [12,13].

While the threats from climate change to low-lying activities like farming are recognised [14,15], current research primarily focuses on the effect of, and adaptation to, rapid-onset hazards such as intensified storm surges and flooding. Examples of investigation into sudden-onset threats in New Zealand include assessments of the impact of storm surges on built assets nationally [16], analysis of associated economic risks of storm surges [17,18], the impact of permanent inundation and coastal erosion hazards [19], the impact of storm surges on stormwater and wastewater networks [20] and associated risks of storm surges and flooding on insurance retreat [21].

By comparison, little has been investigated in relation to the impact of slow-onset climate change-driven hazards, such as rising groundwater. As exceptions, refs. [22,23] map salinity exposure of municipal assets and Chambers et al. (2023) model and quantify uncertainties of the temporal disposition of groundwater inundation [24]. Otherwise, most attention to slow-onset hazards in New Zealand relates to adaptation and management strategies, focussing on either a specific sector of adaptation (e.g., insurance or managed retreat). Harker et al. (2016) discuss managed retreat in New Zealand of housing and associated vulnerabilities [25]. Cuendet et al. (2020) analyse adaptation options for sea level rise in Seaview Gracefield (a low-lying coastal industrial area near Wellington), New Zealand, as a case study [26], and Mourot et al. (2022) develop a methodological framework to support climate adaptation [27].

Otherwise, investigation of slow-onset threats has been so limited globally that the United Nations Climate Change Conference (2012) considers, monitoring over the long term of slow-onset processes has thus far not been adequate in most countries [28]. In the case of sea level rise-driven groundwater rise, most attention has tended to focus on its desirability as a means to buffer water shortages and protect cultures in groundwater-dependent areas [29,30]. However, rises in groundwater are likely to also generate negative impacts. In lowlands, rising groundwater has the capacity to worsen the impact of sudden-onset hazards like flooding [12] and may also impact water quality, including by contributing to salinity through saline intrusion of aquifers [31,32]. Thus, increasingly shallow groundwater may threaten agriculture, necessitating structural adjustments, including potentially adjusting production practices, establishing new infrastructure and/or relocating activities [24,33]. Actions require planning and action at a collective level to facilitate the transformation of food and agricultural production systems [34,35].

In this paper, we use New Zealand—an island nation in Oceania which is highly dependent on export income from farming—as a case study to explore how climate change-driven sea level rise may affect low-lying coastal farming by the year 2100. These lowland coastal areas are often naturally prone to waterlogging and flooding but have been made agriculturally productive through drainage of wetlands, channelisation of watercourses and stop-banking of floodplains [34]. The study integrates the consequences of both slow-onset hazards (groundwater rise) and sudden-onset events (intensified flooding and storm surges) resulting from climate change as they, along with gradual land subsidence, challenge the water management capacity of existing infrastructure. These physical impacts are translated into economic terms, to consider their implications for land-use planning.

Our exploration is based on a conceptual model of a stylised coastal dairy farm confronted by change and the implications for profitability of alternative farm-level responses. The responses explored were informed by consultations with subject-matter experts, with costs based on averages from case studies. Costs reflect the general absence of farm subsidies in New Zealand and the lack of large urban populations and industrial production that could fund major infrastructural solutions such as barrage gates.

2. Materials and Methods

2.1. Define a Stylized New Zealand-Based, Low-Lying, Coastal Dairy Farm as a Case Study

Climate change-driven sea level rise is posing threats to New Zealand by increasing the risk of sudden-onset extreme weather events, such as flooding, as well as by presenting slow-onset threats such as rising groundwater. Understanding the extent of these threats is complicated by the dynamic nature of New Zealand’s land. For example, some of New Zealand’s coastline is subsiding at rates >2 mm/year. This vertical land movement makes a significant contribution to relative sea level rise projections for all scenarios up to 2150 [36]. Additionally, reclaimed drained coastal peatland soils are expected to subside at rates of ~8 ± 4.5 mm/y [37]. Such subsidence is expected to exacerbate existing sudden-onset hazards such as flooding and coastal erosion, while raising the risk of slow-onset hazards such as shallow groundwater emergence and the salinization of wetlands and aquifers [12,38,39,40,41]. Numerous response options for this situation exist (Appendix A).

Here, we focus on the implications of sea level rise for lowland farming in New Zealand. Dairy farming is a major industry in fertile lowlands and floodplains that are prone to waterlogging and flooding. We conceptualise a stylised dairy farm (founded on norms and average traits, scale, type and geophysical settings around New Zealand) within a managed drainage area facing climate change-driven rises in groundwater and intensification of flooding.

We draw on published information on production values and the costs of responses to generate a high-level illustration of the nature of hazard impacts on pasture production over time and the implications for land use. We consider the change in situation from today through to 2100 (75 years) and explore the implications over time.

Associated site adaptation options and practical solutions have been developed in consultation with local subject-matter experts (e.g., dairy consultants), drawing on their real-world, hands-on experience and the costs associated with the documentation they suggested or provided. We aim to maximise the accuracy of our representation of New Zealand dairy farming practices without exploring sensitive information, so we provide generalities for farming. Based on this, we develop a conceptual framework that integrates physical hazards and farm-level practice information with adaptation options to illustrate how the financial future may look. Recognising that this is a conceptual framework based on generalities, it is important to recognise that this framework is not yet suitable for predictive purposes as site-level finances would be needed for that purpose.

For our stylised farm we assume:

• A rain-fed (i.e., non-irrigated) pastoral dairy farm at the national average size of 162 hectares, with the national average stock rate of 2.76 dairy cows/ha [42,43] and an average operating profit of NZD$3300/ha. This is based on average national operating profits reported over the previous seven years [44].
• A farm located in an area already at risk of periodic river flooding and where the local/regional council already operates some form of flood protection (e.g., stop banks) or drainage system from which people benefit.
• Generally flat land such that the distance from groundwater to the surface of the land is similar across the farm.
• A farm operating on land that initially responds well to drainage interventions. Our cost and operating profit assumptions are based on this.

2.2. Consider Slow Onset Threats—Rising Groundwater

Rising sea levels along the coasts are expected to drive groundwater upwards [45]. The extent to which sea level rises are reflected in rising groundwater is uncertain, as it depends on the relative head and flux of freshwater and the presence of surface drainage [46]. Ideally, models would be available to quantify groundwater rise by evaluating the interaction between sea level rise and the local hydrological system. In the absence of this, a conservative assumption (known as the Hydrostatic Rise approach) is commonly used, where the rate of groundwater rise is assumed to be equal to sea level rise [47]. We apply this simplified assumption for our model. Nevertheless, in practice, the relationship between groundwater rise and sea level rise will vary in response to site differences in conductivity, surface and subsurface drainage, recharge, hydraulic gradients, and the presence of agricultural activities [48,49].

For the purpose of this exploration, we draw on work by Hamlington et al. (2024) [6] who estimate sea level rise in the order of 4.5 mm per year, and we assume groundwater rises at the same rate (a sea level rise-to-groundwater rise ratio of 1:1). While the true rate of rise may be higher or lower, the logic of change remains the same. We apply a sensitivity analysis (Section 3.3) to explore implications if the rate of groundwater rise is less or greater than sea level rise.

For the purpose of illustration, we consider implications assuming the following scenarios:

• Groundwater starting at 1 m (GW1) below the ground surface of coastal land and rising at 4.5 mm per year to sit at around 0.7 metres from the surface by year 2100, without intervention.
• Groundwater starting at 0.7 m below the ground surface of the coastal land rising at 4.5 mm per year, to sit around 0.4 metres from the surface by year 2100, without intervention.

Potential Impact on Operational Profits

The effect of rising groundwater on pasture production will depend on a variety of factors including antecedent soil moisture conditions. Under dry conditions (e.g., drought), shallow groundwater can enable improved pasture production growth. However, under wet conditions, shallow groundwater can cause or exacerbate waterlogging, harming pasture production directly as well as exacerbating soil quality issues such as pugging and soil compaction. Yields of perennial ryegrass typically decline as a result of waterlogging [50] and profitability can be expected to fall (Figure 1).

Perennial ryegrass is the most widely sown grass in New Zealand and a feed crop commonly used in dairy farming [51]. Reasonably, we expect that ryegrass pasture productivity will decline as water tables become shallower, until conditions are unsuitable for production. This is conceptualised as a general decline in earnings over time as waterlogging increases (the general downward trend presented in Figure 2). To estimate the effect of rising groundwater on our farm, we draw on Snow et al. (2025) [52] who quantify the relationship between dry matter production for perennial ryegrass and groundwater depth. (See Appendix B for detailed information about the equation).

Reductions in ryegrass dry matter production are assumed to be offset by purchased feed. We assume that the greater the reduction in ryegrass production, the higher the purchased feed requirements and the lower the operational profits. This is logical and representative of the realities in New Zealand. We use proportional declines in dry matter yield arising from waterlogging to infer the potential increases in feed costs required to maintain farm production and the consequent reduction in operational profits that arises. Cost information on feed is adapted from DairyNZ [53].

2.3. Consider Sudden-Onset Threats—River Flooding

Flooding in New Zealand is a common hazard. Between 1968 and 2017, more than 80 damaging floods were reported to have occurred [54]. With increases in sea levels and associated rises in groundwater, the risk of storm surges and river flooding can be expected to increase [12,55,56,57]. The result can be expected to lead to increased costs from extreme events and, reflecting those increasing costs, falling profitability (Figure 1). For the purpose of illustration, we then consider how matters might play out if productivity falls because of sea level rise-driven increases in flood risk over time.

We draw on Paulik et al. (2021) who report the impact of severe river flooding on farms in the Bay of Plenty region [58]. To estimate the effect of increased flood risk on farming, we could apply average annual losses to illustrate the risks of flooding over time (e.g., [59,60,61]). However, average annual losses mask the shocks that farmers experience from individual extreme events and that provide explicit imperatives for the uptake of adaptation. Accordingly, we present the case where a severe flood hits the farm every 10 years to illustrate how climate change leads to spikes in costs (dips in operating profits) at the farm. At a conceptual level, layering these periodic drops in income atop revenues that are steadily declining due to waterlogging results in farm financial performance resembling a ‘ratcheting’ downward trend over time (Figure 2). We assume regular local-level flooding in between these severe events.

We assume a 10-year flooding farm-level cost of NZD$70,000 per event, which includes the transfer of livestock for off-farm grazing during recovery [58]. The intensity of the flooding is expected to increase as climate change progresses [62,63]. While any year could be selected, we assume flooding commences in year 9 to allow us to understand potential profitability over time (including that before flooding-based hardship sets in).

Based on the Clausius–Clapeyron relationship [64], we assume that for every 1 °C increase in temperature, the atmosphere can hold about 7 per cent more water vapour, potentially leading to a similar increase in the intensity of extreme rainfall events (e.g., [65,66]) and increasing costs accordingly. This scenario is indicative only, as actual changes in flood frequency, severity and costs will differ depending on local hydrology and climatology and the specific level and circumstances of the flooding impact. Therefore, as with all our assumptions, these values are for illustrative purposes only.

2.4. Consider Adaptation Actions

At a basic level, it is possible that farmers may not respond to rising groundwater and intensified flood threats. Continuing without change—doing nothing—provides a baseline against which changes in practice can be compared. However, in practice, most farmers can be expected to respond to threats using micro- or macro-level changes to their practices. The Food and Agriculture Organization (FAO) distinguish three broad approaches for agricultural response:

• Approaches that enhance the absorptive capacity of systems to manage negative events using predetermined coping responses in order to preserve and restore essential structures and functions. Actions under this approach enable farming systems to cope with the impacts of a shock in the short run.
• Approaches that enhance the adaptive capacity of systems to adjust or modify so they can moderate harm and/or benefit from opportunities, in order to continue functioning without major qualitative changes. Actions under this approach enable farming systems to better cope with climate change over the medium run through incremental change.
• Approaches that fundamentally change systems (transformational capacity). Actions under this approach involve long-term structural or systematic change, such as developing new production systems or investing in institutional change [67].

Within these approaches, there are a wide number of responses that farmers might adopt. Because we do not know the condition of a farm to begin with, or which action farmers might choose, we take a ‘what if’ approach to assess the implications of climate change on lowland farming. We present a sample of adaptation actions that variously contribute to farming absorptive, adaptive or transformational capacity and assess them using a cost–benefit framework. The actions we consider are as follows.

2.4.1. Absorptive Capacity Responses

Responses that can mitigate the effects of flooding and waterlogging under the absorptive capacity of a dairy farm system include use of standoff pads or well-drained/higher elevation paddocks to avoid grazing in waterlogged paddocks and ‘wintering off’ of livestock (see below). In instances where waterlogged paddocks are grazed either through mismanagement or lack of alternatives, damaged pasture and soil can be rehabilitated via resting, mechanical remediation, or renovation and reseeding, depending on the severity of the damage. We focus on wintering off for the purposes of this study.

Wintering Off

Wintering off involves removing the stock from the dairy farm when they are not milking. Most New Zealand dairy farms aim to have all cows calving within an 8-week period in early spring, milked from spring to late summer or autumn, and then ‘dried off’ (non-lactating phase) in late autumn or early winter. This leaves a period of about 2.5 months in the wetter winter period when the farm can be de-stocked to reduce treading damage and rest the pasture. This response presumes that farmers have or can lease land elsewhere to hold their cattle during winter and early spring. This option also incurs costs to transport cattle on and off the farm, and additional feed costs to sustain them during a period of low pasture productivity. Wintering off does not mitigate progressive waterlogging of soil but can help to maintain productivity as pasture rests.

Paulik et al. (2021) surveyed farmers in the Bay of Plenty to understand costs incurred to transport graze stock elsewhere after flooding (

Table 1. Off-site grazing costs in NZD$. Source [58].

Stock Size	Transport Costs	Grazing Off-Farm Costs
114–642	6000–60,000	9000–112,000

) [58]. As this assessment related to flooding, costs were also estimated for veterinary fees as well as supplementary feed following a flood. We exclude veterinary costs but draw on the transport and off-farm grazing costs, updating them to 2024 values to generate illustrative costs per head of NZD$61 to transport stock off-site and NZD$91 to graze them off-site for four months. For obvious reasons, we assume that wintering off avoids harm to livestock that would ordinarily occur during flooding, reducing associated costs.

2.4.2. Adaptive Capacity Responses

‘Holding the Line’ Through Drainage

Flood control and pumped drainage in New Zealand is often provided as a public service. From a climate hazard perspective, this action mitigates/reduces flood risk. It may also have some benefits related to drainage from waterlogging, albeit limited. We consider the value of flood control and drainage to manage flood costs for our stylized farm. Here we envisage holding the line through higher levels of drainage service. Typically, drainage management structures such as stop banks and associated drainage systems, flood ways and spillways work together as part of a coordinated system that is managed by councils as public assets with operational costs charged back to those landowners that benefit from the infrastructure via council rates. With strict land consenting regulations, farmers operating in flood-managed/drainage areas can have comparatively limited options to add significant structures beyond the public infrastructure already provided by the council, except by installing feeder ditches or subsurface drains flowing to the main drains operated by councils [68,69,70].

We present the case where farmers do not establish their own stop banks, drains and pumps to drain land, but rather lobby for higher levels of service from councils, resulting in higher farm costs via additional targeted rates. The scale of rates for council drainage services is determined based on the land size, proportion of land affected by drainage need (e.g., the area of land requiring pump services) and the class category (Appendix C).

We explore the impact of council drainage services when (i) farms upgrade from no/Class B services to Land Class A services, (ii) farms upgrade to Land Class A services and pumping of 25 per cent of land yearly, and (iii) farms upgrade to Land Class A services and pumping of 25 per cent of land, with land pumped increasing by 15 per cent more every 25 years.

Because there can be substantial variation in land class within a farm, not all land may require drainage (or has Class B drainage) in the first instance. However:

• Over time, sea level rise exacerbates fluvial flood risk, so land may require higher drainage services (Class A drainage) to prevent overflow.
• Ongoing sea level rise can potentially increase river stage such that Class A does not prevent overflow, so pumping is required as well. In such cases, we assume pump services are required for 25 per cent of land every year to help recover from regular flooding.

Establishing Drainage Ditches

Ditches can drain excess water from soil and lower the water table. In this exploration, we focus on the use of drainage ditches as a means to mitigate waterlogging from rising groundwater. Depending on the farm, excavation of ditches may be undertaken using contractors or by the farmers themselves. When established by farmers, the costs of ditches principally relate to labour, although diggers and the transportation of materials will incur fuel costs and potentially equipment hire as well. In addition, the land area occupied by ditches reduces the amount the land that would otherwise be used for grazing.

In practice, fencing and periodic mechanical clearance is recommended for drainage ditches in dairy pastures in New Zealand, with edges vegetated to minimise erosion and overland transport of nutrients and pathogens to waterways [71]. We assume a 1.5 m deep × 1 m wide at base ditch with 2.12 m for fencing and 2.12 m² per metre length of vegetation on each side of the ditch to minimise the risk of collapse. To estimate the magnitude of costs for drainage, we draw on:

• Research conducted on farms to manage diffuse pollution [72], which applies to a generic earth excavation value of NZD$10/m³ inclusive for earth moving, equipment transport and labour.
• The Waikato River Authority’s estimates of fencing at NZD$9.20 per metre for 3-wire electric fencing in the Waikato in 2022 [73] and adjust for inflation to assume a cost of $10 per metre. We assume fencing is established on both sides of the drainage ditch and lasts 20 years (this is a general assumption as in some areas, fencing may only be needed on one side of paddocks and some areas may need it on both sides).
• Data from the Waikato River Authority [73] indicates fence installation costs (e.g., labour, land preparation, and transportation) to be in the order of NZD$5815.38/km in 2022. We update these costs to 2024 values (Appendix D) although values will vary according to landscape, any pre-existing access to equipment and availability of ‘free’ labour (e.g., farmer labour).

We assume that drainage ditches are a permanent infrastructure so there is no lifespan of the ditches. However, ditches require regular mechanical clearance (e.g., of sediment or weeds) to retain effectiveness [71]. We assume that 10 per cent of ditches are re-excavated every three years.

Establishing Water-Tolerant Pasture

We explore the implications of converting pasture from ryegrass to water-tolerant tall fescue grass in the areas that are impacted by high groundwater. This is because tall fescue grass is considered more tolerant of wet soils than perennial ryegrass (and consequently more productive in wet seasons as well) [74,75,76,77,78]. We consider the use of tall fescue in three cases:

• Tall fescue is planted as part of standard pasture renewal today.
• Tall fescue is planted as part of standard pasture renewal in 20 years’ time.
• Ryegrass is actively removed and replaced at cost by tall fescue.

The use of water-tolerant grass such as tall fescue is assumed to have no mitigation impact on flood risk.

In previous studies, greater reduction in growth were identified in roots and shoots under waterlogging conditions for perennial ryegrass and cockfoot grasses than for tall fescue [79,80].

No information was found on the quantified impact of progressive waterlogging on tall fescue that would allow comparison with waterlogging impacts on perennial ryegrass. That said, it can reasonably be assumed that even more water-tolerant tall fescue grass has growth limits when facing waterlogging (see, for example, [80,81]).

For the purpose of exploration, we consider the possible impact of shallow groundwater on tall fescue if losses from waterlogging were five per cent lower than might be expected for perennial ryegrass. In practice, the impacts of waterlogging are variable, so this assumption is only illustrative, and more research is required on this topic. We explore the impact of introducing tall fescue as part of standard replenishment of pasture (i.e., at no additional costs) today or later in the future and if pasture was overhauled (replaced at cost). The use of water-tolerant grasses such as tall fescue can impact profitability; the operational profitability of tall fescue is estimated (see Appendix E).

Retiring Part of the Land to Restore Wetland Vegetation

Ongoing groundwater rise would reasonably be expected to ultimately saturate grassland and, unimpeded, could lead to land eventually transforming to a form of wetland. In these cases, farmers could maintain production in the areas least affected, while retiring more threatened land for development into wetlands. Those (wet)lands would remain exposed to flooding and increasingly shallow groundwater. Wetland development provides the opportunity for farmers to support ecosystem services (such as development of biodiversity and enhanced coastal protection) or benefit from cropping of wetland vegetation via activities such as paludiculture [82]. The type of wetlands emerging as a result of rising groundwater will depend on a variety of factors such as distance to the coast, land slope, inundation level, salinity level, etc. Wetlands might be freshwater in nature (e.g., swamp or marsh) or brackish (e.g., salt marsh, mangroves, or seagrass meadow). We consider a restored coastal wetland to explore implications.

In our exploration, we assume that conversion of land to wetlands takes the form of conversion to salt marsh or mangroves and occurs progressively over the years. We assume 15 per cent of land is converted to begin with, increasing by a further 15 per cent every 20 years:

• Today: 15% of land is converted.
• Yr 20: 30% of land is converted.
• Yr 40: 45% of land is converted.
• Yr 60: 60% of land is converted.

We also explore the implications of converting land over time to mangroves together with introducing water-tolerant grass (tall fescue).

The costs to establish or restore wetlands from pasture around New Zealand vary according to site, type of wetland developed and the purpose of the wetland (e.g., to restore natural ecosystems or processes in an area compared to constructing artificial wetlands to absorb farm runoff). In New Zealand, costs to restore woody, wetland, and other vegetation species in ephemeral wetlands in the southern North Island using voluntary labour for planting and weed management were around NZD$4000/ha in 2012 [83]. More recently:

• Consultations conducted by NIWA to assess the costs of wetland construction in New Zealand [72] revealed landscaping (vegetating) for constructed wetlands for farm management purposes to be around NZD$48,000/ha (personal communication).
• Salt marsh restoration in the Bay of Plenty is reported to be between NZD$20,000 and NZD$190,000/ha, with an average cost of NZD$30,000/ha, not including the cost of the land [84], citing personal communication with the Bay of Plenty Regional Council. These costs are indicative prices estimated based on common approaches for planting (e.g., average speed of planting), labour conditions (e.g., volunteer or paid workers (hourly rates generated using New Zealand’s minimum wage).
• More recently theestimated average cost of mangrove restoration in New Zealand has been estimated to be around USD$52,000/ha [85].

To determine establishment costs for salt marsh restoration, we draw on estimates in the range of NZD$30,000/ha [84] to NZD$48,000/ha [72] to assume indicative establishment costs for salt marsh restoration today of NZD$40,000/ha (see cited research for more details). In both cases, wetlands were physically established and therefore cover all costs including engineering (e.g., coststo reconnect to tidal waterways and provide conditions conducive to restoration). As we used secondary data, our cost of site adaptation can be maintained as conceptual. To determine establishment costs for mangrove restoration, we convert value estimates to present-day New Zealand dollar values [85].

Recognising that wetlands will require maintenance in the early days to ensure that they do not become invaded by weeds, we estimate the annual maintenance costs of wetlands to be around 4 per cent of the establishment costs [83]. We assume annual maintenance costs of 4 per cent of the total establishment cost per year for the first five years after establishment.

We assume all of the necessary site adaptation to foster wetland establishment has been implemented and so the associated costs (e.g., cost of tidal reconnection and appropriate inundation management) are excluded from our model (as these costs vary depending on different levels of impacted conditions).

Potential Benefits of Restored Non-Tidal or Coastal Wetland

The New Zealand emissions trading scheme (ETS) trades carbon emission for offsets at a current carbon sequestration price of NZD$64 per tonne [86]. Coastal blue carbon habitats (salt marsh, mangroves, and seagrass) can sequester carbon within their habitats. To estimate the potential cumulated carbon sequestration rates of wetlands, we assume:

• Cumulative carbon sequestration rates of 0.64 tC/ha/y for restored salt marsh and 0.89 tC/ha/y for mangrove [84].
• A total of 3.67 tonnes of CO₂ stored in the wetland can generate 1 tonne of accumulated carbon [87,88].

The resulting value of carbon sequestration from wetlands is then approximately NZD$202/ha/y for salt marsh and NZD$209/ha/y for mangrove (2024 values). We assume that costs to restore non-tidal (teal) freshwater wetlands would be similar but C sequestration rates somewhat lower [89].

Other than carbon sequestration, wetlands are associated with a wide range of other ecosystem benefits including, for example, coastal protection, soil erosion control, wild fish and other natural aquatic biomass provision services, retention and breakdown of nutrients, habitat maintenance, and spiritual, artistic, and symbolic services [90,91]. Internationally, reported values for these different components of ecosystem services across the world can be found in datasets such as the Ecosystem Services Valuation Database [91] or The Economics of Ecosystems and Biodiversity database (https://www.es-partnership.org/esvd/esvd-download/original-teeb-database/, accessed on 25 November 2025). However, studies that have been tailored to specifically assess the economic value of wetlands in New Zealand are relatively scarce. The estimated economic values f Whangamarino Wetland vary between NZD$60,000 [92] and NZD$55 million [93], depending on the service conferred (

Table 2. Economic values of Whangamarino Wetland. Source: [92,93]. Costs and values are estimated and presented in NZD$ in 2024 values.

Service	Reported Economic Estimate	Source
Replacement cost for flood management	NZD$16,000,000	[92]
Gamebird hunting	NZD$60,000 p.a.	[92]
Existence/intrinsic value	NZD$55,291,671 p.a. 1	[93]

¹ Then converted to present values using the Reserve Bank of New Zealand inflation calculator (https://www.rbnz.govt.nz/monetary-policy/about-monetary-policy/inflation-calculator, accessed on 15 December 2025).

). The gross use value delivered by 166,000 ha of wetland ecosystems was estimated to be NZD$8720 million (2012 terms)—or around NZD$52,530/ha [94]. These values are high compared to average international values estimated for wetlands [95].

Based on a 1988 value of USD$ 15,230,769 p.a. converted to 1988 New Zealand dollars [USD$:NZD in 1988 was approximately 1:1.4333 https://www.poundsterlinglive.com/bank-of-england-spot/historical-spot-exchange-rates/usd/USD-to-NZD-1988, accessed on 8 September 2025].

Representing the economic value of wetlands in this work, including values for both carbon sequestration and other ecosystem services, is precarious. It brings the risk of double counting and overstating the value that wetlands confer. To minimise this risk, we select a single ecosystem service value whose relationship with carbon sequestration is limited—wetland biomass as a source of food production supporting biodiversity ecosystem services. As some risk of overstating wetland values may still exist, we also include a sensitivity analysis of the performance of wetland-related options (Section 3.3).

The values assumed for biomass as a source of food production for species to support biodiversity vary depending on the wetland type involved. For salt marsh, the value of wetland biomass as a source of food production is assumed to be in the order of NZD$3076/ha/year and for mangroves, NZD$7134/ha/year (2024 values; [95]). Recognising that these values only reflect a proportion of the ecosystem services provided by wetlands implies a somewhat conservative assessment of wetland values.

Acknowledging that the biodiversity value of wetlands increases as they mature, we use the formulae on the relationship between restored salt marsh growth and species richness, and we use growth curves of temperate mangrove in Australia as proxies for mangrove maturity and biodiversity [96].

2.4.3. Transformative Capacity Responses

An option to ease financial pressure and reap upfront benefits from current land value might be to sell or lease out some land, while retaining conventional practices on remaining pasture.

Sale of Land for Lifestyle Blocks

Conversion of rural agricultural land to ‘lifestyle blocks’ (small rural holdings primarily used for residential and recreational purposes) has increased rapidly over the last 25 years around New Zealand [97]. As in other developed countries around the world, this has been driven by the demand for lifestyle living (clean air, privacy, peace and quiet, open space, and an enhanced quality of life) and the financial gains from rural subdivision. Commonly varying from 1 to 20 ha with an average size of 5 ha, the number of lifestyle blocks has increased more than 7-fold in New Zealand in the last 25 years [97].

Land near the coast can attract a premium price, while flood risk and elevated water tables may render this land less favourable, resulting in planning and regulatory challenges and reduced value. We therefore assume that only a portion of farmland might be sold off (e.g., the land with desirable locations) and explore implications of a sale of 15 per cent of the land. The mean price of bare lifestyle blocks in 2024 was NZ$287,082/ha [David Shaw, consultancy via this research, conducted on the 25 August 2025 [98], but we assume this would include additional fencing, accessway and subdivision costs. The mean sale price for dairy farmland in 2024 was approximately NZD$48,325/ha [98]. We draw on Dairy NZ [99] to apply a value of NZD$ 42,628/ha, which is the decadal average of sale values for dairy land during 2013–2014.

Lease of Land for Photovoltaic Power

Solar photovoltaic (PV) systems are of increasing interest as a means to support farming land use in New Zealand [100,101]. Land use for this purpose in New Zealand could theoretically take the form of a direct replacement of land use or involve targeting both energy production and farming together (‘agrivoltaics’—e.g., [101]). Theoretically over 80 per cent of agricultural land in New Zealand—around 10 million hectares—should be ‘fairly suitable’ or ‘good’ for agrivoltaic land use [102].

To maintain comparability with land conversion to wetlands, we assume that the lease of land for PV power in our exploration occurs progressively over the years. We assume 15 per cent of land is converted to begin with, increasing by a further 15 per cent every 20 years:

• Today: 15% of land is converted.
• Yr 20: 30% of land is converted.
• Yr 40: 45% of land is converted.
• Yr 60: 60% of land is converted.

A recent assessment of farm-level adoption of agrivoltaic production [100,103] considered the case of a 235-hectare flat-land dairy farm in New Zealand with a total of 860 cows, where 2 ha could be converted to an agrivoltaic production system. It was observed that the PV equipment could be elevated to enable cattle to continue to move across land, allowing grazing and power production to occur simultaneously (‘solar grazing’), albeit at a lower level of grazing (as equipment takes up some of the space). The power generated by the 2 ha of equipment could theoretically offset the costs of PV establishment and reduce grazing by supplementing the power supply (e.g., to support the milking shed). Based on a farm-level estimate of costs and benefits, it was estimated that:

• An agrivoltaic system would be expected to lower financial returns to the dairy farmer compared to conventional farming.
• Due to the relatively high return to farmers of dairy farming, agrivoltaic approaches are likely to be better suited to non-productive dairy areas or might be better suited to the installation of panels on shed roofs, rather than agrivoltaic production per se [100,103].

On the other hand, with the prospect of increased flooding and/or waterlogging from rising groundwater, the value proposition for photovoltaic solar energy production might well rise. In practice, the form of agrivoltaic systems will vary from site to site, according to the type of equipment used, the intensity of placement (space between equipment) and the lay of the land.

In comparison to farm-level adoption of agrivoltaics, Miller (2020) considers the scope for utility-level PV power production, focussing on the suitability of high-producing exotic grassland, low-producing grassland and depleted grassland for utility-scale PV systems [104]. The financial and economic efficiency of utility-scale production would in practice be influenced by access to the broader electrical power grid/distribution system [104].

The alternative to the incorporation of PV systems into farming might be for farmers to lease land for PV power generation. The capital costs for utility-scale solar systems in New Zealand are decreasing and are ‘now close to a point where rate of return becomes acceptable to consider building such a plant’ [104].

Reported solar leasing rates for farms in New Zealand have previously been between NZD$2000 to $6500/ha per year [105,106], anecdotally as far as NZD$8000/ha per year [107]. Properties within 5 km of a power substation suitable for distributing the solar power are likely to be easier to develop than those further away [106]. This makes them more attractive to a developer and likely to command a higher lease price [106]. For the purpose of this study, we explore the implications of leasing 10 per cent of the land to solar power at a price of NZD$4000/ha.

Reflecting the information provided, we explore 13 responses which comprise either singular actions or combinations of actions. In relation to the FAO’s absorptive, adaptive or transformational strategies for agriculture, we explore the responses under four scenarios (Appendix F).

2.5. Estimating Economic Impacts

We assess the nature of economic effects on the farm over time with and without changes. We estimate the gross values (earnings) of the farm by considering the potential earnings with different responses, less the financial costs of operating those responses. Assessments of values such as costs are estimated and presented in New Zealand dollars and in 2024 values. To compare the relative value of different responses, we calculate the net present value (NPV—see Appendix G) of options over time.

Because the impacts of different responses in the face of rising groundwater and flood threats occur over different periods of time, we account for the value of time by using discount rates. Simply put, positive discount rates emphasise values (benefits or costs) associated with an action today compared to those in the future. High discount rates diminish the value of costs or benefits occurring in the future. Low rates, by contrast, assign greater weight to future wellbeing. The public sector investments are subject to a discount rate of 2 per cent for the first 30 years, 1.5 per cent for years 31 to 100, and 1 per cent for year 101 onwards [108]. Alternatively, private discount rates can be approximated using lending rates for private investments. Nationally, private sector loan rates vary, starting from around 4 or 5 per cent [109,110], depending on the value and type of financing undertaken.

For the purpose of our exploration, we apply a discount rate of 5 per cent to farm earnings. We conduct a sensitivity analysis using the national public sector discount rate of 1.5 per cent.

3. Results

We first describe the impact of rising groundwater and increasing flood threat on values of the farm over time without any adaptive response. We then compare the gross values that might be achieved with different responses. Then we consider the net present values of different responses.

3.1. Gross Values Without Change in Management

Without a change in farm management, climate change threatens farm productivity by raising groundwater and increasing the incidence of waterlogging, which can depress pasture production over time. Compounding this, regular and increasingly intense flooding due to climate change can incur costs and depress productivity, causing dips in earnings. In our example, this appears figuratively as a downward ratcheting trend in earnings. In Figure 3 with deeper groundwater, the effect of increasingly intense regular (10 yearly) sudden-onset flooding events is visibly prominent in causing gradual reductions in farm profitability, generating a ’ratchet’ effect. The non-discounted earnings curve is highest, as discounting depresses the value of earnings over time (presenting as lower solid and dashed curves). The higher private discount rate causes the decline to fall more rapidly. Nevertheless, while the impact of climate change (especially from rising groundwater) means that earnings fall steadily over time, the farm continues to earn profits by 2100 (albeit narrowly).

With shallower groundwater, waterlogging resulting from increasingly shallow groundwater depresses profitability more rapidly, compounding the cost of flooding and having an increasingly more powerful influence on farm productivity (Figure 4). Together, climate changed-induced increases in flood severity and shallow groundwater result in lower earnings over time than when groundwater is deeper. The farm faces losses before the end of the century. With discounting, future losses are diminished so the farm continues to face losses before the end of the century but at a diminished rate.

3.2. Gross Values with Change

Regardless of groundwater levels or discounting, responses that include some form of land conversion to mangroves generate the highest gross benefits (

Table 3. Gross benefits in NZD$ (discounted).

		GW = 1000 DR = 5%	GW = 700 DR = 5%	GW = 1000 DR = 1.5%	GW = 700 DR = 1.5%
1	Without intervention	5,717,885	2,953,884	11,772,004	4,756,297
2	Wintering off	4,362,463	1,598,462	8,748,402	1,732,695
3	Council upgrades from none/Class B to Land Class A	5,608,498	2,844,497	11,566,366	4,550,659
4	Council upgrades and pumping	5,368,375	2,604,374	11,029,015	4,013,308
5	Council upgrades and increasing extents of pumping	5,318,944	2,554,943	10,779,651	3,763,944
6	Vegetated, fenced drainage ditches, etc.	6,232,614	3,833,860	13,977,163	8,609,193
7	Water-tolerant grass planted today as part of standard pasture renewal	5,427,662	2,802,650	11,172,048	4,509,017
8	Water-tolerant grass planted in 20 years’ time as part of standard pasture renewal	5,618,528	2,914,876	11,421,582	4,654,534
9	Replacing pasture with water-tolerant grass at cost	5,148,212	2,523,198	10,892,598	4,229,567
10	Progressive conversion to salt marsh	4,465,110	2,376,144	11,460,710	6,882,212
11	Progressive conversion to mangroves	38,316,957	36,227,992	251,988,764	247,410,267
12	Progressive conversion to mangroves plus water-tolerant grass	38,052,304	36,068,387	251,511,343	247,163,038
13	Sale of land to developers	26,079,681	23,730,280	57,491,499	51,528,148
14	Lease of land for PV power	7,494,274	5,405,309	17,640,932	13,062,435

). The highest performer in terms of gross benefits is the conversion of progressively more land to mangroves, followed by progressively converting land to mangroves in tandem with the introduction of water-tolerant pasture. Detailed tables can be found in Appendix H.

The strong performance of the mangrove-related responses is explained by the comparatively high monetized value of environmental services associated with mangroves. This compares starkly with the performance of converting 10 per cent of land to salt marsh: this response performs worst or at least very poorly compared to other responses, even compared to doing nothing. This is explained by the relatively high cost of land conversion and the relatively low value of environmental services assumed for salt marsh.

Based on the exploration conducted, only wetland-related options confer benefits to the public. Due to the monetised values assumed for ecosystem services of wetlands (NZD$3076/ha/year for salt marsh and NZD$7134/ha/year for mangroves—Section 2.4.2), mangrove-related adaptation options offer the greatest benefits to the public (

Table 4. Monetised value of ecosystem benefits (carbon sequestration and food production in support of biodiversity ecosystem services) in NZD$ (discounted; discount rate at 5%).

Adaptation Option	Total Value over Time
Salt marsh	1,784,310
Mangrove	38,793,141

; Figure 5 and Figure 6).

Upgrading council services to include improved flood protection and progressively higher levels of pumping over time also performs poorly. This might be explained by the fact that improved flood protection reduces the costs of flooding but does not mitigate increasing waterlogging of pasture on which farm profitability relies. Also performing poorly compared to other options are the introduction of drainage ditches—where the cost of establishment and ongoing maintenance take time to recover—and wintering off—which offers no impact on waterlogging and only occasional and partial respite from flood costs.

3.3. Net Benefits of Change

The net benefits of responses—the NPV—are calculated by comparing discounted impacts of response to discounted farm performance if no response was undertaken—and then subtracting the costs. Discounting is influential here because it diminishes the value of benefits or costs occurring in the future. Benefits that take time to accrue in the future diminish more rapidly with a high discount rate but diminish more slowly with a lower discount rate. The result is that wetland-related actions that take time to generate mature benefits perform better economically with the higher private discount rate as benefits diminish less.

While discount rates influence benefits, they also reduce the economic value associated with costs. This means that the economic performance of options that might otherwise be poor (due to, say, high upfront or ongoing costs) are estimated to be better with higher discount rates. For example, in this exploration, the NPVs of wintering off, improved services from councils and planting water-tolerant grass today are estimated to be negative over time on account of the relatively low savings achieved compared to earnings without action. The higher discount rate ‘burns off’ the losses faster than the lower discount rate. As a result, these options fare better economically with a high discount rate, although they still result in losses relative to the status quo.

Overall, mangrove-related actions are estimated to offer the highest NPVs, followed by the transformation of land from agriculture to lifestyle blocks and/or PV power (Appendix H). These latter actions benefit from immediate upfront benefits. The weakest performing responses were estimated to be wintering off (which offers relatively low benefits to groundwater threats), conversion to salt marsh (which has relatively low environmental benefits for the upfront costs) and actively overhauling pasture (replanting at cost, rather than as part of standard pasture renewal; this has sizeable upfront costs which are difficult to recoup). The ranking of responses remains largely the same as the ranks for gross benefits (Appendix H).

3.4. Sensitivity Analysis

Two sensitivity analyses were conducted. First, as information on the exact value of ecosystem service benefits from wetland-related options was not available, it is possible that the potential benefits of wetlands may be over- (or even under) estimated. Accordingly, a sensitivity analysis is provided of the performance of wetlands if monetized environmental benefits (carbon sequestration and value of food production in support of biodiversity ecosystem services) are excluded. In this case, wetland-associated options perform poorest out of all options because (i) the costs of wetland establishment are high and (ii) profit-generating grazing land would be surrendered for non-monetized environmental purposes (

Table 5. Gross benefits of wetland-related options in NZD$ (discounted).

		GW = 1000 DR = 5%	GW = 700 DR = 5%	GW = 1000 DR = 1.5%	GW = 700 DR = 1.5%
10	Progressive conversion to salt marsh	2,680,800	591,834	5,197,574	619,077
11	Progressive conversion to mangroves	−476,185	−2,565,150	−315,946	−4,894,444
12	Progressive conversion to mangroves plus water-tolerant grass	−740,837	−2,724,753	−793,368	−5,141,673

and

Table 6. Net benefits of wetland-related options in NZD$ (discounted).

		GW = 1000 DR = 5%	GW = 700 DR = 5%	GW = 1000 DR = 1.5%	GW = 700 DR = 1.5%
10	Progressive conversion to salt marsh	−3,037,085	−2,362,050	−6,574,430	−4,137,220
11	Progressive conversion to mangroves	−6,194,069	−5,519,034	−12,087,950	−9,650,741
12	Progressive conversion to mangroves plus water-tolerant grass	−6,458,722	−5,678,638	−12,565,371	−9,897,970

).

Additionally, a sensitivity analysis is provided of the performance of adaptation options under different rates of groundwater rise. As discussed in Section 2.1, the rate of groundwater rise may not necessarily be the same as that of sea level rise. For example, sea level rise-driven rises in groundwater can slow as the distance from the coast increases.

Accordingly, a sensitivity analysis was conducted to assess performance where groundwater rise is half the speed of sea level rise and where groundwater rise occurs at 150 per cent the speed of sea level rise (sea level rise-to-groundwater rise ratios of 1:0.5 and 1:1.5 respectively). In these cases, mangrove-related options remain the highest performers of adaptation options, provided that assumed ecosystem benefits are monetised (

Table 7. Gross benefits in NZD$ (discounted)—sea level rise-to-groundwater rise ratio of 1:0.5.

		GW = 1000 DR = 5	GW = 700 DR = 5	GW = 1000 DR = 1.5	GW = 700 DR = 1.5
1	Without intervention	6,044,116	3,521,826	13,047,975	7,164,053
2	Wintering off	4,688,694	2,166,404	10,024,373	4,140,451
3	Council upgrades from none/Class B to Land Class A	5,934,729	3,412,439	12,842,337	6,958,415
4	Council upgrades and pumping	5,694,606	3,172,316	12,304,986	6,421,064
5	Council upgrades and increasing extents of pumping	5,645,175	3,122,885	12,055,622	6,171,700
6	Vegetated, fenced drainage ditches, etc.	6,250,201	3,875,591	14,016,520	8,702,578
7	Water-tolerant grass planted today as part of standard pasture renewal	5,736,781	3,340,978	12,382,206	6,793,347
8	Water-tolerant grass planted in 20 years’ time as part of standard pasture renewal	5,932,214	3,460,808	12,638,385	6,949,871
9	Replacing pasture with water-tolerant grass at cost	5,457,331	3,061,528	12,102,756	6,513,897
10	Progressive conversion to salt marsh	4,682,490	2,747,582	12,191,002	8,210,750
11	Progressive conversion to mangroves	38,534,338	36,599,429	252,719,056	248,738,805
12	Progressive conversion to mangroves plus water-tolerant grass	38,258,211	36,420,334	252,203,796	248,423,143
13	Sale of land to developers	26,356,977	24,213,031	58,576,075	53,574,741
14	Lease of land for PV power	7,711,655	5,776,746	18,371,224	14,390,972
	Alternative wetland options with non-monetised ecosystem service benefits:
10a	Progressive conversion to salt marsh	2,898,180	963,272	5,927,866	1,947,615
11a	Progressive conversion to mangroves	−258,804	−2,193,712	414,346	−3,565,906
12a	Progressive conversion to mangroves plus water-tolerant grass	−534,930	−2,372,808	−100,915	−3,881,567

and

Table 8. Net benefits in NZD$ (discounted)—sea level rise-to-groundwater rise ratio of 1:0.5.

		GW = 1000 DR = 5	GW = 700 DR = 5	GW = 1000 DR = 1.5	GW = 700 DR = 1.5
1	Without intervention	−1,355,422	−1,355,422	−3,023,602	−3,023,602
2	Wintering off	−109,387	−109,387	−205,638	−205,638
3	Council upgrades from none/Class B to Land Class A	−349,510	−349,510	−742,989	−742,989
4	Council upgrades and pumping	−349,510	−349,510	−742,989	−742,989
5	Council upgrades and increasing extents of pumping	206,085	353,765	968,545	1,538,524
6	Vegetated, fenced drainage ditches, etc.	−307,335	−180,848	−665,769	−370,706
7	Water-tolerant grass planted today as part of standard pasture renewal	−111,901	−61,018	−409,590	−214,182
8	Water-tolerant grass planted in 20 years’ time as part of standard pasture renewal	−586,785	−460,298	−945,219	−650,156
9	Replacing pasture with water-tolerant grass at cost	−1,361,625	−774,244	−856,973	1,046,697
10	Progressive conversion to salt marsh	32,490,222	33,077,603	239,671,081	241,574,752
11	Progressive conversion to mangroves	32,214,095	32,898,508	239,155,820	241,259,090
12	Progressive conversion to mangroves plus water-tolerant grass	20,312,861	20,691,205	45,528,100	46,410,688
13	Sale of land to developers	1,667,539	2,254,920	5,323,249	7,226,919
14	Lease of land for PV power	−1,355,422	−1,355,422	−3,023,602	−3,023,602
	Alternative wetland options with non-monetised ecosystem service benefits:
10a	Progressive conversion to salt marsh	−3,145,935	−2,558,554	−7,120,109	−5,216,439
11a	Progressive conversion to mangroves	−6,302,920	−5,715,538	−12,633,629	−10,729,959
12a	Progressive conversion to mangroves plus water-tolerant grass	−6,579,046	−5,894,633	−13,148,890	−11,045,620

). Overall, changes in the rate of groundwater rise have only marginal impacts on economic payoffs. With slower groundwater rise, waterlogging harm to pasture is slowed and the basic operating profits improve both with and without adaptation. The improved value of gross benefits without action marginally lowers the payoffs of adapting. However, the general ranking of adaptation options remains the same. The reverse situation occurs where the rate of groundwater rise is faster (lowering basic operating profits due to higher harm to pasture and improving investment payoffs from adapting). Detailed figures for this sensitivity analysis are provided in Appendix I and Appendix J.

As with sensitivity analysis of the ecosystem service values, wetland-related options perform poorest if wetland ecosystem values are not monetised.

4. Discussion

Not adapting to climate change implies a steady decline in farm profits over time as increased flood risk and rising groundwater hamper operations. In our exploration, sudden-onset decadal flooding has an important influence on farming when groundwater is deeper (1000 mm from the surface), creating noticeable economic shocks to the farm [111]. However, as sea level rise pushes groundwater closer to the surface, the slow-onset hazard of waterlogging can begin to bite. With both hazards, farming might still make some level of operating profit by 2100 even if no changes are made. However, profits are noticeably lower, especially where groundwater is shallow and a private discount rate applies. Our exploration indicates that the economic performance of a farm could improve if changes are made. In this exercise, responses that include some degree of mangrove restoration generate potentially high benefits due to the monetised values assigned to the (public) ecosystem services they confer. Compared to other responses, mangrove-related responses in our exploration consistently offer the highest potential net benefits to our farm, regardless of groundwater levels or discount rates. Combining some mangrove conversion with water-tolerant grass generally offers the highest net benefits. However, the majority of the benefits associated with mangrove-related options would not accrue to the farmer/landowner unless they can be monetised. If the (public) ecosystem service benefits of wetlands—especially mangroves—are not realised (monetised), these adaptation options perform very poorly indeed.

The positive performance that mangroves with monetised ecosystem service benefits offers does not mean that mangrove-related responses are necessarily practical across New Zealand. First, although salt marshes occur in suitable habitats throughout New Zealand, mangroves only grow in the northern half of the North Island of New Zealand [84]. Secondly, rapid mangrove expansion in many New Zealand estuaries has been driven by increased deposition of fine sediments from anthropogenic land management practices. Thirdly, mangrove or salt marshes are not included in the current Emission Trading Scheme. This expansion has led to reduced views of open water and obstruction of recreational access, thereby contributing to negative public perceptions of mangroves—despite their well-established ecological value [112]. In contrast, salt marsh has significantly declined in extent (due in part to mangrove expansion and coastal squeeze) and provides a lower-growing and more diverse ecosystem. Consequently, its restoration is generally perceived more positively and is actively underway in some areas around New Zealand, e.g., [113]. However, achieving suitable conditions for its establishment can be challenging [113,114] and may require special management such as tidal control structures [115].

Additionally, the economic performance of responses that include wetlands depends on the value assigned for the environmental services they provide. While we applied a comparatively modest ecosystem service value for mangroves (NZD$7,134/ha for mangrove food production to support biodiversity), the value could be different. Stewart-Sinclair et al. (submitted) indicate that, depending on the specific ecosystem service considered, the global average value of ecosystem function of mangroves could be as low as NZD$3/ha (in that case, for the role of mangroves in regulating water flows) [95]. If this value is used, the economic performance of mangrove-related responses would fall extensively, and the rank of these options would be very low.

As well, our assumed ecosystem value for the role of salt marsh in food production to support biodiversity is low compared to others [95,116]. Not surprisingly, this resulted in the relatively poor economic performance of the salt marsh-related response. These points underscore the ongoing importance of assessing, recognising and confirming the economic value of alternative ecosystem service values for New Zealand’s wetlands.

Furthermore, the economic value assumed for mangroves in our framework reflects impacts that are commonly public in nature—that is, they accrue to wider society rather than specifically to the farmer. As a result, the incentive for farmers to consider investing in ecosystem-based responses to climate change threats may be small because the benefits they provide are currently perceived as externalities so they cannot reap most of the benefits of their efforts. Unless the farmer can monetise the (public) benefits of mangrove restoration (e.g., via subsidy payments, cost sharing or ecosystem service markets), they might more logically sell off land for development, as the boost in income today results in the highest NPV outside of mangrove restoration options.

Given the environmental, social and aesthetic values that wetlands can confer, the protection and restoration of wetlands across New Zealand have been an area of considerable interest in recent years (see, for example, [92,117,118,119,120]). If New Zealand envisages a future that benefits from healthy wetland ecosystems and the services they provide, there needs to be a means for farmers to reap the public benefits of their investments. This is because our exploration indicates that using The Treasury discount rate for public investments (1.5 per cent) [108] could result in an economic (primary public) benefit for our farm that is six times higher than what could have been reaped with a higher, private discount rate. This could be used towards dialogue for cost sharing with farmers.

Potentially, cost sharing for wetland restoration might be introduced to subsidise farmer investments in wetlands in the public interest. Cost sharing might theoretically begin by extending existing trading systems in New Zealand. Presently, wetland ecosystems do not form part of New Zealand’s emissions trading scheme (ETS). Including wetlands could enable farmers to recoup the costs of wetland establishment or protection via their carbon sequestration functions. The real value of wetlands such as mangroves—and the real potential in cost sharing for their establishment and protection—may thus be in benefits conferred outside of conventional trading systems in New Zealand altogether.

An alternative strategy might be to consider subsidised loans that more closely reflect the social discount rate of public investments or to support the development of markets where commercial wetland products could be generated. The opportunities for paludiculture in the Waikato region of New Zealand, such as the development and sale of wetland crops for bioenergy, water filtration and/or for crafting, but these are unlikely to be economically viable under current conditions [82]. Viable responses need to be suitable for specific farm-level situations and for different waterlogging conditions with different farming traditions, and the adaptation practices would also need to be integrated specifically. With limited time and resources, these details are excluded in this research.

In addition to mangrove-related responses, our exploration highlights the potential value of other transformative responses. Following mangrove-related responses, actions involving transforming some land to use for PV power or development are the next economically valuable options. In practice, challenges may exist for such transformative options. First, transformation of productive land for land development would require the handover of prime land which is most desirable. This would imply that the remaining land is less productive and profitability for the remaining farmland would be lower, potentially harming the future of the farm. Furthermore, consenting requirements for new land use could be demanding. While this is not a challenge for farmers who sell their land, it could be time consuming and demanding for developers and buyers, impacting the viability of this option. For example, there may be objections to changes in land use, such as possible resistance to the aesthetic/environmental amenity impact of land used for PV power [101,105]. It may also be difficult to resume farming later, if farmers wished to do so at the end of their lease [105]. In any event, mangrove-related responses are not necessarily practical for New Zealand at present as noted above.

In any case, conversion of land to some alternative uses may only offer short-term financial relief. Transfer of land for development may offer income in the short term but it will only have a specific window of opportunity [121]. While areas of high groundwater and flood risk are not good for farming, they are also unlikely to be appealing to developers without significant earthwork operations, which would increase costs and could exacerbate downstream flooding risks. Landowners would need to pre-empt critical water table levels in order to benefit from sales. The same higher, less at-risk land that farmers are mostly likely to wish to retain is likely to be the same land they may be most able to sell. In selling off—or even leasing out—the most productive land to other users, farmers will be left with less productive land for ongoing farm operations. As a result, their ability to generate profits over time might actually decline faster with this response. It should be recognised that timing will also matter for other responses. For example, planting of water-tolerant grasses may buy time for farmers but continually rising water tables are likely to mean that even water-tolerant pasture has a finite window of use.

5. Limitations of This Study

Our framework and story are based on a stylised New Zealand temperate farmland situation. Farm operations and responses are simplified to present general cases. However, variations across farms will affect the size of the window for adaptation. For example, debt levels and management approaches vary, influencing the capacity of farms to absorb climate change-induced shocks over time. Likewise, differences in site conditions will impact farm resilience. For example, for farms located on coastal lowlands with minimal slope (e.g., θ ≤ 0.5%) [122], the window for maintaining farming systems in the face of increasing risk of flooding and shallow groundwater will shrink faster than for other farms. Potentially, responses to threats such as cut-and-carry systems and water-tolerant grasses will be less viable in these areas in the short to medium term and farmers will have fewer choices. In addition, our assumptions are that our archetypical farm is all relatively flat and prone to sea level rise and thus all the land is at some risk of flooding. In the real world, this land may not be permitted for house development because of the flooding risk and the structural risks due to building on waterlogged soils. Costs to build foundations in these conditions would be higher and insurance more expensive and/or difficult to find. Furthermore, low-cost housing is unlikely to provide optimal financial returns. Development in these low-lying lands is likely to favour the less at-risk, higher elevation, more productive land.

This study identifies the depth of groundwater as a critical factor in how long farmers have to plan ahead. Where groundwater is presently deeper than 2 m, farmers have a larger window of adaptation opportunity than where groundwater is shallow. With a rise in groundwater of 4.5 mm per year, land with groundwater starting 2 m from the surface could remain relatively unaffected for decades. As a result, time remains for farmers to contemplate future responses such as drainage or implementing cut-and-carry.

By comparison, farms with already shallow groundwater have fewer years to adjust practices. They will be increasingly vulnerable to intensifying flooding as well as groundwater rise. Their ability to recover from the flood events with increasingly moist soils is likely to become increasingly difficult.

While these simplifications exist, our framework could be extended to accommodate farm system absorptive and adaptive capacity by coupling it to a farm system model and running simulations on a daily timestep.

We apply functions to estimate the magnitude of possible losses to pasture from incremental waterlogging [52]. In practice, the quantity of harvestable dry matter affected will be subject to specific soil and groundwater characteristics and farm management such as drainage, fertiliser usage and the grazing regime. Our estimates of loss are therefore illustrative.

Due to data gaps, we do not include the impacts of climate change on salinisation of soil in this exploration. As sea level rises, it is likely that soil will be exposed to greater salinity, and ryegrass pasture production could be expected to decline more rapidly than the value estimated only by reference to groundwater heights. Where sea level rise does cause soil salinisation [123], profitability can be expected to decline more rapidly. The logic of our exploration remains the same at this point, but the windows for change in adaptation would contract more rapidly. In this case, farmers would likely also have incentives to explore alternative adaptation options (e.g., more salt-tolerant pasture).

Our exploration assumes there are no catastrophic failures in existing stop banks and drainage systems that drastically change the viability of farming. We also assume a static ongoing sea level rise rate of 4.5 mm/year, with no future acceleration or abrupt rises (e.g., an abrupt ice shelf collapse scenario). This may not be the case given the increased rates observed over recent decades and the potential for acceleration as a result of tipping points and cascading effects not accounted for in sea level rise models [124]. We also acknowledge that in coastal aquifers, the relationship between groundwater rise and sea level rise depends on conductivity, surface and subsurface drainage, recharge, hydraulic gradients, and the presence of agricultural activities [48,49]. Thus, the rate of groundwater rise may not be linear with the rate of sea level rise. The projections to 2100 should therefore be viewed as an exploration of the method and as a conceptual framework rather than a precise scenario or a prediction because the research assumes the groundwater rise is the same rate as the sea level rise.

To tease out the broader story of possible adaptation implications, some of our assumptions are simplistic. For example, we assess the value of drainage and of drainage plus pumping, on the assumption that they both maintain production levels in the face of climate change. In practice, these options would have variable impacts on the farm. (A farmer would not need to consider the drainage and pumping option if the council drainage scheme alone prevented flooding.) However, these assumptions are used to enable us to consider wider implications.

Finally, our exploration of responses that involve wetland development does not include the cost to reconnect inundated land with the coast. If no human interventions are involved (e.g., setting up a drainage management system to enforce the reconnection), the potential time frame of the natural reconnection process is also not included. In this research, we made reasonable assumptions that the reconnection, salinity level, water level, and other environmental conditions will be suitable for coastal wetland to be established, noting that the associated human intervention cost varies depending on specific characteristics and may not be feasible is some situations.

6. Conclusions

Farming the lowlands of New Zealand will be impacted by climate change through rising groundwater and increasing intensity of flooding. A variety of possible responses exist to adapt to these slow- and sudden-onset threats, but the extent to which they will be valuable will depend on the unique conditions of farms, not least the present extent of shallow groundwater and the severity of flood risk. As the productivity of coastal lowlands become increasingly compromised by climate change, the need to consider alternative uses of lowlands—and cost-sharing arrangements to enable them—will become critical.

Our exploration also highlights the role of sudden- compared to slow-onset threats in the profile of risk to farmers. Farmers may be less able to focus on production between flood events and increasingly focus on chronic waterlogging as rising groundwater impedes flood recovery. In a world dominated by flood reports [125,126,127,128] and the fear of worsened floods [129], climate change will bring slow-onset threats such as groundwater rise into stark focus in New Zealand’s lowlands and highlight the need for long-term planning.