Climate change is predicted to alter the current pest status of Globodera pallida and G. rostochiensis in the United Kingdom

Abstract The potato cyst nematodes Globodera pallida and G. rostochiensis are economically important plant pathogens causing losses to UK potato harvests estimated at £50 m/ year. Implications of climate change on their future pest status have not been fully considered. Here, we report growth of female G. pallida and G. rostochiensis over the range 15 to 25°C. Females per plant and their fecundity declined progressively with temperatures above 17.5°C for G. pallida, whilst females per plant were optimal between 17.5 and 22.5°C for G. rostochiensis. Relative reproductive success with temperature was confirmed on two potato cultivars infected with either species at 15, 22.5 and 25°C. The reduced reproductive success of G. pallida at 22.5°C relative to 15°C was also recorded for a further seven host cultivars studied. The differences in optimal temperatures for reproductive success may relate to known differences in the altitude of their regions of origin in the Andes. Exposure of G. pallida to a diurnal temperature stress for one week during female growth significantly suppressed subsequent growth for one week at 17.5°C but had no effect on G. rostochiensis. However, after two weeks of recovery, female size was not significantly different from that for the control treatment. Future soil temperatures were simulated for medium‐ and high‐emission scenarios and combined with nematode growth data to project future implications of climate change for the two species. Increased soil temperatures associated with climate change may reduce the pest status of G. pallida but benefit G. rostochiensis especially in the southern United Kingdom. We conclude that plant breeders may be able to exploit the thermal limits of G. pallida by developing potato cultivars able to grow under future warm summer conditions. Existing widely deployed resistance to G. rostochiensis is an important characteristic to retain for new potato cultivars.


| INTRODUCTION
Climate change has the potential to alter the distribution of animals, but outcomes can vary. In the absence of habitat management, it may result in future extinction of some species, as reported for drought-sensitive butterflies in the United Kingdom (Oliver et al., 2015). Worldwide, both vertebrate and invertebrate species have moved towards higher latitudes over a circa 25-year period (Chen, Hill, Ohlemuller, Roy, & Thomas, 2011;Hickling, Roy, Hill, Fox, & Thomas, 2006). Data for many of the 612 crop pests and pathogens analysed established a global move poleward since the 1960s for some organisms but not for either Globodera pallida or G. rostochiensis in the Northern Hemisphere (Bebber, Ramotowski, & Gurr, 2013).
Both these species of potato cyst nematodes (PCN) occur throughout the potato-growing regions of the United Kingdom (Minnis et al., 2002)  ). There are few cultivars grown widely with high levels of resistance particularly to G. pallida. Early success in breeding resistance to G. rostochiensis was achieved with a single gene from Solanum tuberosum ssp. andigena. It has provided durable, qualitative resistance to this nematode in cultivars such as Maris Piper. Breeding for resistance to the more common forms of G. pallida is more complex. No single gene offers complete resistance to all populations of this nematode which vary in the level of virulence they offer to partially resistant cultivars (Dalton, Griffin, Gallagher, de Vetten, & Milbourne, 2013). Rotational control is an important pest management strategy for PCN that counters their reproductive success on a host plant by allowing natural decline rates when other crops are grown (http://pota toes.ahdb.org.uk/online-toolbox/pcn-calculator).
PCN are host-specific parasites that co-evolved over 15-21 9 10 6 years with wild potato species (Solanum L. section Petota Dumort.) of which 130 species are recognized in Peru and Bolivia (Spooner & Hijmans, 2001). The climate of the Andean highlands preadapted both the two PCN species and their host potato plants to cool-temperate climates worldwide where the crop is now grown.
G. pallida is adapted to high altitudes and is considered to have undergone an expansion northwards within Peru in the Miocene as the Andean chain rose in that region. Phylogenetic analysis of G. pallida populations has been used as a molecular clock to determine when an altitude threshold of 2.0-2.5 km was reached for the elevating Andes in different regions of Peru. G. rostochiensis is assumed to originate from where uplift of the paleo Andes was less extreme, and therefore, the climate is slightly warmer (Plantard et al., 2008).
The females of both Globodera species retain all eggs within a cyst formed by tanning of their body walls at death. There is normally a single generation per potato crop, and the encysted eggs remain dormant until infective juveniles hatch from them in response to root diffusate from potato plants (Forrest & Farrer, 1983;Perry & Beane, 1982). A partial second generation has been observed for some populations of G. rostochiensis in the United Kingdom (Evans, 1969;Jones, 1950) and Italy (Greco, Brandonisio, Tirro, & de Marinis, 1988). Potatoes are grown widely in England and Scotland with planting of main crops from mid-April and harvest up to early October (Daccache et al., 2012;Gregory & Marshall, 2012 (Whitehead, 1992) and mid-June for early planted potatoes in Belgium (Ebrahimi, Viaene, Demeulemeester, & Moens, 2014).
The difference in the altitude adaptation of the two PCN species in the Andes, together with previous work, suggested a comparative approach for this study to define if their likely responses differ to the future increase in summer temperatures that have been projected for the United Kingdom (Jenkins et al., 2009;Parker, Legg, & Folland, 1992;Trenberth et al., 2007). Being soil borne, nematodes respond to soil rather than air temperature. As soil temperatures are rarely recorded and are not outputted from global climate models, modelling studies tend to use air instead of soil temperatures. Garc ıa-Su arez and Butler (2006) showed for three sites in Ireland that over the last century annual mean soil temperatures increased more than air temperature, and rises and falls do not occur at the same time. To represent recent and future soil temperatures for 10 sites across the UK potatogrowing area ( Figure 1, Table S1), we validated a soil temperature model (SoilClim) (Hlavinka et al., 2011)  hatched, infective juveniles (J2s) of each species were added to the soil after three weeks of plant growth. Juveniles were hatched from G. pallida (pathotype Pa 2/3) or G. rostochiensis (pathotype Ro1) eggs within cysts at 20°C using root diffusate collected from three-weekold potato roots. J2s were washed four times in tap water and pipetted into the soil at three locations around the planted tuber at a density of one juvenile per lL water. Immediately after soil infestation, the potato plants were transferred to heat mats set at 15, 17.5, 20, 22.5 and 25°C in a glasshouse with a 16-hr day length.
At least three replicate plants per temperature were grown for each time point. Soil temperature in each pot was monitored using an iButton (Maxim Integrated, San Jose, California, USA) and was within AE1°C of the set mat temperature throughout the experiment. Females were collected from roots at weekly intervals, from their first appearance on the root surface at three weeks until nine weeks postinfection by washing them through a series of 1000-, 150-and 63-lm sieves.
Images of females were taken using a Leica MZ16 stereo-binocular microscope and a MicroPublisher 3.3 RTV colour camera (QImaging, Surrey, BC, Canada). Projected surface area was measured in mm 2 using Image-Pro Analyzer 7.0 (Media Cybernetics Inc., Rockville, MD, USA). The eggs within some newly formed cysts of both species were counted after measuring their projected surface area to provide a calibration curve that relates area to egg number.
2.2 | Population growth of Globodera on a range of cultivars at 15, 22.5 and 25°C Tubers of nine potato cultivars widely grown in the United Kingdom (Solanum tuberosum L. var. Arsenal, Cara, Desiree, Estima, Innovator, Markies, Melody, Maris Peer and Maris Piper) were planted and grown as described above with three or four replicates per temperature for each potato cultivar in soil containing G. pallida (pathotype Pa 2/3) at a density of 5 eggs g À1 . The plants were grown on heat mats set at 15, 22.5 (all cultivars) and 25°C (Desiree and Maris Peer only) in a glasshouse with a 16-hr day length. After twelve weeks of growth, plants and soil were allowed to dry. Roots and soil were mixed together, and three 100 g samples were collected from each pot. Egg and cyst counts were carried out using standard procedures by an agricultural extension company (ADAS UK Ltd, Wolverhampton, UK). Cysts were recovered from dried soil samples using a Fenwick can, they were opened and the number of eggs was quantified on a counting slide (see Southey 1986 for details). The same experiment was performed for G. rostochiensis with potato cultivars Desiree and Maris Peer.

G. rostochiensis
Eight Solanum tuberosum L. var. Desiree plants were grown for two weeks before 1,500 J2 G. rostochiensis/plant were introduced to each as described above. The temperature was 21.2 AE 1°C throughout the experiment. Half the plants were selected at random and harvested after nine weeks. The remaining plants were harvested at 16 weeks. The number of cysts/100 g soil, the projected surface of each cyst and their egg content were measured as described earlier.

| Dependency of development rate of
Globodera on temperature Temperature-dependent development of Globodera females over time was investigated using the Gompertz model as modified by Zwietering, Jongenburger, Rombouts, and van't Riet (1990) to give the parameters a biological meaning: F I G U R E 1 Locations of selected 5 9 5 km weather grid cells .
where l m is the maximum specific growth rate, that is the tangent in the inflection point; k is the lag time before first egg production (x-axis intercept of the tangent); and A is the asymptote which is defined as the maximal female surface area achieved.
The lag time was set to a minimum of two weeks as production of embryonated eggs is not expected to occur before this time point.

| Soil temperature simulations
Soil temperature at 10 and 20 cm depth was simulated using Soil-Clim (Hlavinka et al., 2011) which requires daily minimum and maximum air temperature, precipitation, radiation, latitude and altitude of the location as input data. All SoilClim simulations were performed for light and medium soils (Trnka et al., 2014) with notional planting and harvest dates in mid-April and early October (Daccache et al., 2012;Gregory & Marshall, 2012). Soil temperature was simulated for all combinations of a constant (3 t ha À1 ) and variable canopy, with or without irrigation. For the variable canopy, the total amount of biomass cover was increased linearly from 0 to 18 t ha À1 during the initial crop development once it had emerged. It was maintained at 18 t ha À1 during mid-season before a decrease from 18 to 10 t ha À1 in the late season. Irrigation was simulated by maintaining readily available water at a minimum of 40% until increased by rainfall (Daccache, Weatherhead, Stalham, & Knox, 2011 Predictions of future Globodera pressure were calculated using our data for the effect of temperature on female development and population size, together with simulated soil temperature at 10 and 20 cm depth during female development. The potato root system is concentrated in the upper 30 cm of the soil layer (Asfary, Wild, & Harris, 1983), with PCN distribution proportional to root length density (Storey, 1982). Therefore, we used estimates of root length density with depth to weight the simulated soil temperature. Soil layers 0 to 15 cm and 15 to 30 cm were assumed to correspond to Soil-Clim soil temperature estimates at 10 and 20 cm soil depth and were weighted 0.45 and 0.55, respectively (Asfary et al., 1983). The SoilClim simulations were based on a variable canopy with irrigation to provide realistic combinations for conditions experienced by PCN.

| Statistical analyses
All data were analysed using a standard statistical package (SPSS

| RESULTS
3.1 | The effect of temperature on growth of female Globodera The mean projected surface areas of 2,899 and 2,398 collected females of G. rostochiensis and G. pallida respectively were measured over 3-9 weeks after adding juveniles to soil (Fig. S1). Gompertz curves were fitted to the data set with a minimum lag of two weeks before the first females were present on the root surface (minimum value of R 2 = 0.77 except for G. rostochiensis at 25°C where  Figure 3a). Results for the two cultivars were also similar for G. rostochiensis. Both the decline from 22.5 to 25°C and the increase between 15 and 22.5°C were significant for Desiree and Maris Peer (at least p < .01, a priori contrast, one-way ANOVA, Figure 3a). Further analysis indicated that the effects for both species were mainly due to changes in cysts /100 g soil although eggs/ cyst were suppressed for both species on the two cultivars at 25°C (Fig. S2).
The comparative reproductive success of G. pallida at 15 and 22.5°C was studied for a further nine cultivars. The reduction in reproductive success for each cultivar is given in Figure 3b expressed as a percentage for each species at 22.5°C relative to their corresponding means at 15°C. Data collected for Arsenal and Innovator were excluded from the analysis due to low multiplication (<1) on these cultivars at both temperatures. The overall reduction from the mean for the remaining seven cultivars at the higher temperature was 39 AE 4% cysts/100 g soil, 21 AE 4% eggs/cyst and 50 AE 4% eggs/g soil. The reduction from means at 15°C was significant (<0.001 in all three comparisons, multivariate ANOVA, pairwise comparisons, Bonferroni adjustment for multiple comparisons). There were no significant differences among cultivars for number of females or eggs/g soil (one-way ANOVA and SNK, p < .05, Figure 3b).
The reduction in eggs/cyst for Estima at 22.5°C was significantly greater than for Maris Peer, Markies and Maris Piper (SNK, p < .05).

G. rostochiensis
The number of cysts/ 100 g soil was 6.75 AE 1.17 (mean AE SEM) after nine weeks with a significantly higher value of 31.1 AE 6.58 after 16 weeks (p < .001; t test). Given that the number of cysts from the first generation of G. rostochiensis is complete by 7-8 weeks (Fig. S1a), the increase suggests a second generation con- Rothamsted and East Malling. Peaks and troughs of daily soil temperature at 10 cm depth were generally well estimated (Fig. S3a, b).
The goodness of fit between observed and simulated daily soil temperature ranged from R 2 values of 0.73 to 0.82 for Rothamsted and 0.83 to 0.92 for East Malling for different years (Fig. S3c, d). Average air temperature for the potato-growing season from mid-April to early October was predicted to increase from the recent period by a mean of 1.9-2.4°C (depending on the location) by the 2040s with the medium-emission scenario and to 3.9-5.0°C by the 2080s with the high-emission scenario (Fig. S4a). For the same period, total precipitation decreased by a mean of 8 to 31 mm for the 2040s for the medium-emission scenario and by 22 to 58 mm for the 2080s for the high-emission scenario (Fig. S4b). The latter figure represented about 80% of the mean precipitation over the potato-growing season during the recent years.
Daily soil temperatures were generally higher and more variable at 10 cm than 20 cm depth for both the recent and future simulations ( Fig. S5) with a greater effect for the light than the medium soil type (Fig. S6). The medium soil type in combination with variable canopy and irrigation is the most prevalent combination for potato growing in the United Kingdom. With this combination, the mean soil temperature for the medium-emission scenario increases at 10 cm (20 cm) from the recent to the 2040s by 1.9 to 2.7°C (1.8 to 2.6°C) for June at the ten different locations and 2.0 to 2.6°C (2.0 to 2.6°C) for July (  (Table S2). Differences in increases in mean soil temperature varied more depending on the canopy compared to irrigation effects (Table S2). For both June and July, recent median soil temperatures were usually below 15°C for the northern sites and 15°C or slightly above 15°C elsewhere (Figure 5b the results differed significantly when a constant canopy was assumed for SoilClim (Fig. S8a, b). Irrigation compared to no irrigation on the other hand did not change the results significantly ( Fig. S8c). Interannual variability for mean June and July is high and increases in the future (Figure 5b, c) which indicates that some years will have a larger impact on Globodera.

| DISCUSSION
The effect of temperature between 15 and 25°C on female reproductive success differed between G. pallida and G. rostochiensis. Both number of females per plant and final female size and hence number of eggs per plant were reduced progressively for G. pallida at temperatures above 17.5°C (Figure 2 and Fig. S1). In contrast, the F I G U R E 4 Number of observed eggs against that predicted from projected surface area for each cyst collected at nine weeks and 16 weeks from two batches of plants after infection of cv Desiree with hatched juveniles of Globodera rostochiensis. No outliers with low egg content for their size were detected at nine weeks, but seven outliers were present in the cysts recovered at 16 weeks number of G. rostochiensis females developing on potato was only suppressed above 22.5°C with no decrease in female final size over 15-25°C. This differential effect on number of eggs/g soil produced by the two species was also evident from comparing reproductive success at 15, 22.5 and 25°C on cv Desiree and Maris Peer (Figure 3a). The reduction in reproduction of G. pallida was of particular interest in relation to projected future UK summer temperatures and was found to be host independent for seven cultivars studied (Figure 3b).
The reduced number of G. pallida females developing may arise from less efficient root invasion, mortality of developing females or a higher proportion of males in unfavourable conditions as sex is determined by environmental conditions in planta (Perry, Wright, & Chitwood, 2013). It seems unlikely to arise from differences in hatch, as there is no substantial effect for either species over the range studied in the recent work by Kaczmarek, Mackenzie, Kettle, and Blok (2014). Furthermore, number of developing females was also reduced at the higher temperature for G. pallida when hatched J2s were applied directly to the soil (Figure 2). It could relate to the known differential temperature effects on endogenous lipid reserves used for mobility and root invasion by this nonfeeding stage (Robinson, Atkinson, & Perry, 1987). The small, significant reduction in fecundity must occur after sex determination and suggests suboptimal conditions for the feeding female. Our results for female development are consistent with previous work that suggests that G. rostochiensis has a slightly higher thermal optimum (for both number of females and number of eggs per female) compared to G. pallida (Berry, Stone, Parrott, & Edwards, 1977). A higher optimum temperature for G. rostochiensis has also been found for hatch of its infective juveniles (Foot, 1978;Franco, 1979;Kaczmarek et al., 2014;Robinson et al., 1987). Overall, the results from our study and previous studies establish a preference for G. pallida for a lower temperature range compared to G. rostochiensis.
F I G U R E 5 Monthly mean soil temperature (Tsoil) at ten sites as a weighted mean over 10 and 20 cm soil depths for the medium soil type for ( F I G U R E 6 Proportion predicted change for six of the ten sites in the number of eggs per plant using the relationship in Figure 2c and the median values given in Figure 5b, c (mean over June and July) for Globodera pallida (filled bar) and G. rostochiensis (patterned bar). The change to the 2040s is given in dark grey, to the 2060s in medium grey and to the 2080s in light grey. The remaining four sites have median soil temperature below 15°C in recent times which is below the range studied and are therefore not shown The aim of exposing developing females of Globodera to diurnal fluctuations from 17.5°C to up to 32.5°C for one week was to examine the likely effect of short periods of high ambient temperatures. Diurnal fluctuations had a significant effect on the development of growing females of G. pallida measured one week after this heat stress, but some recovery was evident after a further week.
This suggests short periods of high temperature do not suppress multiplication of this species in contrast to sustained high temperatures above about 17.5°C. As previously shown, G. rostochiensis has a higher thermal optima compared to G. pallida and diurnal fluctuations from 17.5 to 32.5°C had no significant effect on the development of growing females for this species at either time point during recovery.
The SoilClim model simulates the recorded soil temperature accurately at East Malling and Rothamsted for all seven years compared (Fig. S3). This suggests that it provides a useful basis for future projections in conjunction with the weather generator that enabled a spatial resolution of 5 9 5 km. This scale is sufficient for estimating regional effects within the United Kingdom. The projected increases in soil temperature during June and July ( Combining the data in Figure 2 with the climate change effects suggests a differential effect on the two species. Multiplication of G. pallida in the six most southern sites is estimated to be reduced by approximately 30%, 40-50% and 50-60% in 2040s, 2050s and 2080s for the high-emission scenario ( Figure 6). Figure 3a indicates that the reduction in the southern sites might be somewhat lower than presented here, but both data agree on a negative trend. In contrast, similar increases in reproductive success are predicted for G. rostochiensis for the same period and conditions but with higher variation between sites. The effect of an increase in mean temperature in the four most northern sites cannot be estimated as current levels were below 15°C and therefore outside the range of the growth experiments. As the medians of future mean soil temperature are between 15 and 20°C, it is anticipated to be insufficient to have either a detrimental effect on G. pallida or to favour G. rostochiensis.
Our results suggest that further work to add a soil temperature Unlike G. pallida, G. rostochiensis maintained its capacity to multiply at 22.5°C (Figures 2 and 3) and completed a generation in 6-7 weeks postinfection of J2 (Fig. S1). A partial if not full second generation was indicated both by the recovery of more cysts at 16 weeks compared to nine weeks postinfection and by the presence of cysts collected at the second time point with low egg content for their size (Figure 4). Some populations of G. rostochiensis in both the United Kingdom (Evans, 1969;Jones, 1950) and Italy (Greco et al., 1988) show a less than complete entry into dormancy of the first generation of eggs and succeed in completing a partial second generation on potato crops. Multiple generations occur for another cyst nematode, Heterodera schachtii on the sugar beet crop, which has a more prolonged growing season than potato plants. Heterodera schachtii can achieve up to five generations per crop in the warm conditions of the Imperial Valley of California but only typically two generations in the cooler soils that prevail in Northern Europe (Thomason & Fife, 1962 (Brown, Towers, Rivington, & Black, 2008;Gregory & Marshall, 2012). If a shift in planting potatoes towards earlier dates occurs in the future, this would place the start of female development into May. The soil temperatures during May would favour G. pallida as they are cooler than in June and July with the median of the mean monthly soil temperature at or below 17.5°C until the 2040s for all sites (Figure 5a). Potato yields in England are predicted to increase from approximately 2.9 to 6.5% by mid-century due to warmer temperatures, assuming current nitrogen management and unconstrained water availability. Current irrigation schemes will not meet needs to achieve future yields in approximately 50% of years with 14 to 30% more water required by mid-century (Daccache et al., 2011). The importance of PCN will be increased if the crop experiences water stress more often as the parasite reduces water acquisition by the root system (Fatemy & Evans, 1986). A shift to the north and west would lessen irrigation demands (Downing et al., 2003), but the effect may be slow because of the investment levels required of successful potato growers (Daccache et al., 2012).
It is generally assumed that PCN was introduced from S. 10 6 years. Some animals are likely to overcome the impact of climate change by range changes (Hof, Levinsky, Ara UJo, & Rahbek, 2011), but this does not apply to G. pallida because it is already present throughout much of the United Kingdom (Minnis et al., 2002).
The prevailing consensus is that climate change normally outpaces microevolution processes that enable the adaptation required to remain at some localities (Hof et al., 2011). Exceptions include Daphnia magna which has a rapid life cycle and lives in shallow pools susceptible to changes in water temperature. The planktonic crustacean showed a 2°C increase in the maximum temperature at which it shows locomotor activity over a two-year period (Geerts et al., 2015). It seems unlikely that Globodera will have a similar capacity to achieve such rapid microevolution given its infrequent reproduction.
Our work suggests dual priorities for potato plant breeders, that is to exploit the thermal limits of G. pallida and continued incorporation of resistance against G. rostochiensis to counter possible benefits to it from warmer temperatures in the United Kingdom.