Scarcity of ecosystem services: an experimental manipulation of declining pollination rates and its economic consequences for agriculture

Pollination system

The study was conducted in four commercial fields producing seeds of Brassica campestris L., chinensis group (pak choi), in the Canterbury Plains, New Zealand. The study was carried out on private land, with permission from the owners. Two fields were producing open-pollinated seeds and two were producing hybrid seed. The genus Brassica (Cruciferae: Brassicaceae) includes many crop species that are commercially important for oil seeds, leaf and stem vegetables, condiments, forage, fodder and green manure (Friend, 1985). For most brassicas, although flowers are hermaphrodite, there is some degree of self-incompatibility and pollination by insects such as bees (Apoidea) or flies is usually required. Although open-pollinated (self-fertlising) crops set seed through a combination of selfing and by pollinators, hybrid (pollinator-dependent) crops, which have separate male and female plants, require pollinators to transfer pollen between them. There are three bee and four fly species, which are the key pollinators of this crop in the study region (Rader et al., 2009). Although, all these species are efficient in transferring pollen, managed honeybees are more effective due to their high population, which leads to higher rates of visitation (Rader et al., 2009).

Experimental plots

In each of the four fields, the crop was sown with a precision drill with an inter-row spacing of 30 cm and with 5 cm between plants in the rows. The two fields with open-pollinated crops were sown on 10 October 2005 and the two hybrid fields were sown on 25 September 2005 for female rows and on 2 October 2005 for male plants, with a female to male ratio of 3:2. The mean size of the fields was c.10 ha. In each, an experimental plot (10 × 10 m) was set up 20 m from the nearest field edge. In each plot there were five treatments arranged in a randomized block design. These comprised plants whose inflorescences were covered with 1 mm white mesh bags for one, two, three or four weeks, or uncovered during the four-week flowering period. Each treatment was replicated six times with a 2 m distance between replicates. For the analysis, the treatments were considered in terms of the proportion of time the plants were exposed to pollinators (pollination rates; 0–no pollinators i.e., bagged for the entire pollination period of 4 weeks; 0.25–bagged for three weeks; 0.50–bagged for 2 weeks; 0.75–bagged for 1 week; and 1.00: no bagging).

All hybrid crops began flowering during the first week of December 2005 and were available for pollination for four weeks. During this period, New Zealand seed producers (arable crops) bring in honeybee hives at a density of 4 hives ha⁻¹ and at a cost of NZD $120/hive/1 month. All the experimental plots in each field were at least 50 m from hives. At the end of flowering, experimental plants were individually removed from the plots, labeled with replicate and field information and placed in paper bags. They were dried in the shade at room temperature for four weeks and the following information was recorded:

• Seeds per pod: Seeds were counted in 50 randomly selected pods for each replicate.

• Seed yield per plant: Seeds for each plant were separately threshed and cleaned using a sieve (2 mm). The seeds were stored in paper bags, which were then weighed individually to calculate seed yield (g dry weight) per plant.

• Proportion of unfertilised pods per plant: For each treatment, the numbers of fertilised and unfertilised pods were recorded.

Data were analysed by randomised-block ANOVA followed by LSD pair-wise comparisons of the two pak choi types (open-pollinated and hybrid) using GENSTAT 7.2. A log₁₀ transformation was used to normalize the variance for seed weight.

Economic impacts

Most flowering plants in New Zealand except grasses require pollinators for pollination and honeybees are the most effective pollinators of commercial crops (Rader et al., 2009). Data for New Zealand on types of crops that require pollinators, their area, total production in tons and market value in NZD (2013) was obtained from the Ministry of Primary Industries, New Zealand website. Data on area and production were not available for vegetable, clover and rye grass seed, however, their market values were available. Therefore, for these three crop types, those values were used in our calculations. Plants’ pollinator-dependency rates and proportion of honeybees as pollinators for each crop were obtained from the literature (Klein et al., 2007; Gallai et al., 2009). We selected a list of 18 crops including fruits, vegetable and pasture seed where honeybees account for 90% of the pollinators. The market value of each crop was then used to calculate the value of pollination services. The economic value attributed to pollination by honeybees was obtained from the following Eq. (1). (1)$TE{V}_{psc}=V_{mc}\times D_{ic}\times P_{hbc}$ Where,

• TEV_psc = Total economic value attributed to pollination services by bees in each crop

• V_mc = Market value of the crop

• D_ic = Insect dependency ratio of the crop

• P_hbc = Proportion of insect pollinators that are honeybees in each crop

Seed yield data in the current study were used to estimate the economic impact of a reduction in the current pollination services in New Zealand due to decreases in pollination rates (a surrogate for a reduction in pollinator numbers) for both hybrid and open-pollinated crops. Percentage decrease in yield was calculated from the seed yield data at each of the four pollination rates.

Changes in the economic value of pollination services provided by honeybees was calculated according to the marginal reduction in yield and economic value due to varying pollination rates by using Eq. 2. (2)$\Delta TE{V}_{psc}=TE{V}_{psc}-(TE{V}_{psc}\times V_{prc})$ Where,

• ΔTEV_psc = Change in the value of pollination services by bees in each crop

• V_prc = Percentage change in the economic value at different pollination rates in each crop

We calculated change in the economic value for each crop at different pollination rates. The aggregated economic value of pollination services provided by honeybees at a national scale was calculated by aggregation of the total value assigned to each crop. Change in total economic value for the main pollinator-reliant crops in New Zealand was also calculated at different pollination rates.

Experimental assessment

Seed yield per plant

There was a significant difference between the open-pollinated and the hybrid crops for seed yield with no manipulation (p = 0.002, Table 1). At lower pollination rates (0 and 0.25), the former produced significantly higher yield than did the hybrids. There was a greater increase in seed yield per plant in the hybrid crop with increasing proportion of time that pollinators had access (Fig. 1) with the maximum increase at 0.50 pollinator exclusion. In open-pollinated crops there was an increase with each 0.25 increase in pollinator access but the rate of increase in seed yield (Fig. 1) was not as high as that found in the hybrid crop. For the latter, there was a reduction of 60% in seed yield compared to 37% in the open-pollinated ones in the complete absence of pollinators (0–Pollination rate). At 0.25 pollination rate, hybrid crops suffered a 52% decline in seed yield whereas open-pollinated showed a 26% decline. At 0.50, hybrids lost 24% and open-pollinated lost 22% of their potential seed yield. At 0.75 pollination rate, hybrids and open-pollinated fared equally, with reductions in yield of 15 and 16%, respectively.

Table 1:

Summary of the results produced by ANOVA for seed weight (g), seeds per pod and percentage of unfertilised pods in two hybrids and two open-pollinated fields.

	Seed weight (g)	Seeds per pod	% Unfertilised pods
Hybrid	1.7 ± 0.3	11.7 ± 1.1	53.6 ± 9.0
Open-pollinated	2.2 ± 0.2	16.08 ± 1.3	40 ± 5.1
Significance	p < 0.001	p < 0.001	p < 0.001

DOI: 10.7717/peerj.2099/table-1

Estimated economic losses from reduced pollination rates—an extrapolation

The total market value of the 18 crop types, to which the experimental results were extrapolated, examined in this study is NZD $2.55 billion annually (Table 2). Based on insect pollinator dependency ratios for each crop and the proportion of pollinators that are honeybees in each case (Klein et al., 2007; Gallai et al., 2009), we estimated the economic value of honeybee pollination services for each crop (Table 2) at varying pollination rates. We used only the data obtained from declines in seed yield from open-pollinated crops in this study, as the 18 crops used in Table 2 are open-pollinated. The total economic value of pollination services provided by honeybees for these 18 crops was NZD $1.96 billion annually. Figure 2 provides the annual total economic value at different pollination rates for these New Zealand crops. The economic loss to New Zealand agriculture could be in the range of NZD $295–728 million/year depending on the extent of future changes in pollination provided by honeybees.

Marginal change in pollination services

Past pollination studies have used cages or bagging to completely exclude pollinators and study the effect on seed yield, seed per pod etc. (Langridge & Goodman, 1975; Williams, 1986; Free, 1993). However, the consequences of varying pollinator rates on yield of these types of experiments have rarely been analysed. It is clear that pollinator populations are in a state of decline in managed landscapes (Kremen, Williams & Thorp, 2002; Potts et al., 2010; Breeze et al., 2014). Also, varroa mite can seriously reduce managed and wild honeybee colonies, imposing serious economic losses (Simpson, 2002). Our results indicate that if pollinators are very rare or absent from a cropped area, open-pollinated crops could provide a better alternative as their decline in yield in the current work was up to 37% compared to hybrids where, the decline could be as high as 60% (Fig. 2). However, hybrid crops are almost always of higher value than open pollinated varieties of the same crop in market. There would thus be a financial loss if a hybrid crop is simply replaced. Such relative ‘improvements’ still do not fully address the serious implications of pollinator losses for New Zealand and world agriculture. Globally, there has been an increasing trend in the cultivation of pollinator-dependent crops since 1961 (Aizen & Harder, 2009; Gilbert, 2016; Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services, 2016). These crops have also produced yield increases by using honeybees (Garibaldi et al., 2009). However, in some parts of the world, especially USA and Europe, recent declines in managed honeybee colonies are serious (Potts et al., 2010; Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services, 2016). The manipulated reduction in exposure to pollinators in the current study serves as an indicator of the potential consequences of future honeybee population declines.

Our logic in this study assumes that a 50% decrease in exposure to pollinators in terms of flowering time is equivalent to a 50% decrease in the availability of managed bee hives in farmland. We argue that if farmers use a fixed number of bee hives to maximise production of crops that depend on pollinators, then any reduction in hive number will impact on yield, as observed in this study (Fig. 1). However, although we manipulated access by pollinators to the experimental crops, it remains unknown whether the ambient pollinators populations could give maximum possible pollination rates in the controls. It follows that if more pollinators had been present, control yields may have been higher and the proportional reductions in pollination rates achieved here would have probably been even higher. Also, the availability and pollinator-dependency of adjacent crop and non-crop plants in the vicinity of the experiment is unknown and that could have reduced the number of pollinators visiting the un-covered, control plots. It is also important to note that the insect dependency rates, i.e., how many pollinators are honey bees, were obtained from global datasets, which also includes several New Zealand based studies (Klein et al., 2007). However, further investigation is required to confirm change in seed yield under varying pollination rates and varying population of pollinators (e.g., by different number of beehives) per field in an experimental setup.

Economic impacts of loss of pollinators

Pollinator declines have been reported worldwide (Williams, 1984; Falk, 1991; Buchmann & Nabhan, 1998; Kevan & Phillips, 2001; Kremen, Williams & Thorp, 2002; Millennium Ecosystem Assessment, 2005; Winfree et al., 2009; Potts et al., 2010). These have rarely resulted in complete failure of crops, but more frequently resulted in reduced yield. The scarcity of pollination services on farmland shifts the supply and demand function in the market (Southwick & Southwick, 1992). As most ES are not substitutable or are substitutable only to an extent, the demand and supply curve for pollination services will resemble Fig. 3, with increased prices to consumers at existing levels of demand, while producers sell fewer products but at higher prices. The supply curves of ES are usually vertical (Costanza et al., 1997; Costanza et al., 2014) as their supply is not affected by the forces of the economy. They are, however, influenced by those economically-driven changes which reduce the biodiversity that provides ES. In the case of pollination services, managed honeybee hives can be increased or decreased in number on farmland to change the extent of this service, so if the supply curve shifts from S₀ to S₁, the price will increase from P₀ to P₁ (Fig. 3). This can have further economic implications for beekeepers and food markets (Barrionuevo, 2007; Teagasc, 2007). Globally, there is a need to grow more food to meet the demand of growing populations and changes in consumption patterns. To help meet these challenges, there is a need to reduce reliance on honeybee colonies to provide pollination. Ecosystems with a high rate of delivery of ES with minimal ecosystem dis-services (Zhang et al., 2007; Wratten et al., 2013) in agricultural landscapes can assure continuous supply of pollination services by wild pollinator species (Kremen et al., 2004; Winfree et al., 2007). For example, favouring wild pollinator species, this can provide continuous supply of pollination services (Kremen et al., 2004; Winfree et al., 2007) thereby providing high rate of delivery of ES with minimal ecosystem dis-services (Wratten et al., 2013).

Modifying existing agricultural systems to enhance ES requires a range of mechanisms such as payments for ES (PES; Wunder, 2005). There are already agri-environment schemes that target enhancement of pollination services due to the recognition of their high economic value (Ricketts et al., 2004; Albrechdt et al., 2007). Adoptions of these schemes depend upon policy changes, as they require financial investment at national and farm level. Set-aside land or reducing the intensity of farming may result in a higher provision of pollination services but only if such practices are informed by sound ecological science (Kleijn et al., 2011). However, a consequence may be that farmers may suffer foregone benefits for example lose some of their crop yields. PES schemes should consider these options, including paying for the foregone income that otherwise would have been generated by farmers (FAO, 2007).