climate change effects on plant-pest-natural enemy ...nature.berkeley.edu › garbelottowp ›...

July 26, 2010 18:42 9.75in x 6.5in b1011-ch11 Handbook of Climate Change and Agroecosystems

Chapter 11

Climate Change Effects on Plant-Pest-NaturalEnemy Interactions

Andrew Paul Gutierrez∗,†, Luigi Ponti∗,‡ and Gianni Gilioli∗,§

∗CASAS Global (Center for the Analysis of Sustainable Agricultural Systems),37 Arlington Ave., Kensington, CA 94707

†Division of Ecosystem Science, University of California,203a Mulford Hall, Berkeley, CA 94720

‡Gruppo “Lotta alla Desertificazione”, ENEA CR-Casaccia,Via Anguillarese 301, 00123 Roma, Italy

§Department of Management of Agricultural and Silvicultural Systems,Mediterranean University of Reggio Calabria,

89060 Reggio di Calabria, Italy

[email protected]

Introduction

Climate change is expected to increase temperatures globally and alter patterns ofrainfall (IPPC, 2004) and other derivative factors that can alter species distribution,abundance and impact in natural, agricultural and medical/veterinary vector/diseasesystems in unknown ways. The term climate encompasses the long-run patternof numerous meteorological factors (e.g., temperatures, humidity, relative humid-ity, atmospheric pressure, winds, rainfall, and others) in a given location or largerregion, while the term weather refers to short-term current measures of these fac-tors. Weather affects the physiology (e.g., development, growth and reproduction,survival, diurnal and seasonal phenology), interactions with other species, and otheraspects of the biology of poikilotherm species (i.e., species unable to regulate bodytemperature) in time and space (see e.g., Andrewartha and Birch, 1954; Larcher,1995; Wellington et al., 1999; Walther, 2002) (Fig. 1). Projected climate changescould modify extant relationships and interactions in food chains and webs (i.e.,plant-herbivore-natural enemy interactions) with unknown consequences in natural,

209


210 A.P. Gutierrez et al.

Fig. 1. A partial list of factors affecting species distribution and abundance.

and agricultural systems, and in medical and veterinary diseases vectored by insects,ticks and other poikilotherm organisms (Gutierrez and Baumgartner, 1984; 2007;Patz et al., 2003; Baylis and Githeco, 2006). The effects of climate change on anytrophic level may cascade to lower and higher levels in food chains and webs affect-ing system regulation (sensu Hairston et al., 1960) and stability.

If the climate change is severe enough, the geographic range and the distributionand abundance of species will be altered in unknown ways leading to potentiallydisruptive effects on trophic and possibly ecosystem structure and function (seeSchreiber and Gutierrez, 1998). Climate change may prove catastrophic for somespecies and human populations unable to adjust to habitat change and unable tomigrate to favorable areas. The increasing introduction of invasive species withglobalization is a further important complication. Current losses due to invasivespecies in agricultural and natural systems are in excess of a $130 billion per yearin the United States alone (Pimentel et al., 2000), and hence they have becomea major focus of much of the climate change research on biological systems (e.g.,alien weeds, insect pests (e.g., aphids, mealybugs, fruit flies, bollworms, defoliators,forest bark beetles), and diseases).

The question of how best to study biological effects of climate change on com-plex biological systems across spatial and temporal scales remains unresolved (sensuHolling, 1992).Among the approaches are: (1) time-series observations to documentbiological responses to changes in climatic variables; (2) remote sensing analysis of


Climate Change Effects on Plant-Pest-Natural Enemy Interactions 211

data; (3) climate envelope approaches (statistically-based ecological niche models(Beaumont et al., 2009) and physiologically-based ecological niche models (Fitz-patrick and Nix, 1968; Gutierrez et al., 1974; Sutherst et al., 1991); (4) physiologi-cally based demographic modeling (Gutierrez et al., 2006a).

Field Observations

Field studies conducted over several years have followed the effects of weather onthe phenology of plant and animal species at both the local and regional levels. Someexamples include bloom dates of apple, grape and lilac (Wolfe et al., 2005), olive(Galan et al., 2005) and other plants (Gordo and Sanz, 2009). In insects, expansionsof geographic range have been predicted in pine processionary moth (Battisti et al.,2005), spruce budworm (Williams and Liebhold, 1997); southern green stink bug(Tougou et al., 2009) and many others. In a six-decade long study, Gordo andSanz (2005) found increasingly earlier time of first appearance for the honey bee;cabbage white butterfly, potato beetle and olive fly (Fig. 2). Gordo and Sanz were

Fig. 2. The time of first appearance of four insect species at Tortosa, Spain, beginning 1943 (modifiedfrom Gordo and Sanz, 2005).



concerned climate change could disrupt the synchrony between host plants andinsect herbivores that in trophic chains and webs evolved to cope with expectedrange of fluctuations in their thermal and other requirements. A major drawback ofthe field survey approach is the length of time required to conduct the study, the lackof generality and explanatory power of the results, and the inability to project theresults to other areas.

Remote Sensing

The coverage and resolution of biophysical data derived from remote sensing (RS)satellite data substantially increase our potential to assess the effects of climatechange on ecosystems on a regional and global scale. Missing variables, low res-olution, inadequate duration, temporal and spatial gaps, and declining coverage inremote sensing data are pervasive (Clark et al., 2001; Pettorelli et al., 2005), but manyof the limitations of RS are being overcome (Turner et al., 2003; Mendelsohn et al.,2007; Camps-Valls et al., 2009). Although, satellite RS data are subject to errors thatsubstantially reduce their ecological applications when not integrated with reliablefield data (Kerr and Ostrovsky, 2003); an integration using new interdisciplinaryapproaches is emerging termed satellite ecology (Muraoka and Koizumi, 2009).

Spectral vegetation indices are widely used to assess ecological responses to envi-ronmental change (Tucker et al., 2005) including the very popular Normalised Dif-ference Vegetation Index (NDVI) that measures canopy greenness (Shippert et al.,1995). Information derivable from NDVI (such as plant phenology, biomass, anddistribution) is of prime importance for terrestrial ecologists (Pettorelli et al., 2005),and novel methods are becoming available for measuring seasonal vegetation pat-terns and dynamics at the landscape level (i.e., landscape phenology) (see Liang andSchwartz, 2009). RS data are being used to map and monitor invasive plant species(see Pengra et al., 2007; DeFries, 2008). NDVI approaches have also been used forlinking vegetation to animal distribution and dynamics (Jepsen et al., 2009).

NDVI and other satellite-derived vegetation indices have been used for moni-toring and early warning of insect herbivore damage in a climate-change contextfor outbreaks of insect defoliators in forests (Eklundh et al., 2009), mortality oflodge pole pine induced by bark beetle (Coops et al., 2009), and especially locustoutbreaks. Examples of the latter include FAO’s pioneering studies on desert locustsin Africa (Hielkema et al., 1986; Cherlet et al., 1991; Cressman, 1996; Voss andDreiser 1997), Australian plague locust (McCulloch and Hunter, 1983), East Asianmigratory locust in China (Ji et al., 2004; Ma et al., 2005; Liu et al., 2008), andlocust infestations in Uzbekistan (Sivanpillai and Latchininsky, 2007).

Prediction of vector-borne diseases is difficult due to the complexity of humanfactors involved, but satellite imaging can play a critical role by monitoring major



environmental drivers (Ford et al., 2009). RiftValley fever (RVF) is a mosquito-borneviral disease in Africa that has been detected by RS (Linthicum et al., 1987). Using acombination of satellite measurements including NDVI,Anyamba et al. (2009) madethe first prospective prediction of an RVF outbreak for the Horn of Africa. Satellitedata are being harnessed to monitor and predict malaria, a mosquito-borne diseasethat kills more than a million people annually (Rogers et al., 2002). Remotely-sensedland surface temperature was used in Italy for early detection and risk assessmentof tick-borne encephalitis (Rizzoli et al., 2007), a zoonotic vector-borne diseasewhose incidence has increased in Central Europe during the last few decades (Carpiet al., 2007). Remote sensing has also been successfully applied to the study of otherarthropod vectors of disease such as blackflies (vectors of onchocerciasis), tsetseflies (sleeping sickness), and sandflies (leishmaniasis) (Kalluri et al., 2007).

Ecological Niche Approach

Ecological Niche Models (ENMs) attempt to characterize and climatically model theecological niche of a species using observed weather and knowledge of its currentdistribution. These models may be either statistical or physiologically based, or beboth. Another term used for this is the climate envelope approach.

Statistical Ecological Niche Models

ENMs developed using statistical methods correlate the current distributions ofspecies based on surveys, museum collections and other records with climate vari-ables. The ENM for a species (i.e., a pest) may then be used to map its potentialgeographic distribution, its invasive potential in new regions, and using climatemodel scenarios to locate the future global areas of suitability (Estrada-Pena, 2008;Mitikka et al., 2008). Beaumont et al. (2009) summarized the assumptions underpin-ning statistical ENMs: they are usually calibrated assuming the current distributionis the best indicator of its climatic requirements; that the distributions are in equi-librium with current climate; and that climate niche conservation is maintained inboth space and time. Major applications of this approach have been to estimate thepotential range of invasive species.

Araujo and Guisan (2006) discuss challenges to statistical ENMs stating that . . .

“In spite of the widespread use . . . (e.g., Guisan and Thuiller, 2005), important con-ceptual, biotic and algorithmic uncertainties still need to be investigated if [these]models are to make important contributions for conservation and bio-geographicalresearch [and agriculture].” Araujo and Guisan reviewed Hutchinson’s (1957) fun-damental and realized niche concepts that are core to ecological niche modeling.They interpret the fundamental niche as defined by the resources and limiting factors



required for species persistence, while the realized niche is defined by the constraintspreventing the exploitation of resources. Inter-specific interactions are essential com-ponents of the realized niche as repeatedly demonstrated in the applied disciplineof biological control (see Rosen and DeBach, 1979). Araujo and Guisan point tothe general absence in ENMs of facilitator species that influence resource utilizationby other species, and further note that niche models explore only snapshot corre-lations that are difficult to use for assessing feedback mechanisms of species withtheir environment (see Davis et al., 1998). Estrada-Pena (2008) concludes ENMs areconceptually unable to distinguish between the transient and equilibrium responseof a species to a stochastically and dynamically changing environment.

Roura-Pascual et al. (2009) compared five techniques for developing statisticalENMs to investigate the potential distribution of Argentine ants in the Iberian Penin-sula: generalized linear models, generalized additive models, generalized boostedmodels, genetic algorithm for rule-set prediction, and maximum entropy models.They found geographic differences in the predictions of the different approaches andin their ability to identify areas of uncertainty regarding the species’ invasive poten-tial. Latimer et al. (2006) showed that making distribution models spatially explicitimproves the accuracy of characterizing the environmental response of species anddrawing inference about species niche relations, distributions, and the effects ofhuman disturbance. Thuiller et al. (2005) developed bioclimatic niche models for96 plant species and subspecies endemic to South Africa that are invasive elsewhere.They posit that if species distribution patterns from the area of origin were available,ENMs could be powerful tools for unbiased first-step screening of areas identifiedas potential invasive sites.

Hickler et al. (2009) studied the effects of changes in water availability on pre-dictions by various ENM species distribution models. They concluded that futurechanges in species ranges and biodiversity should be interpreted with caution, andthat more process-based representations of the water balance of terrestrial ecosys-tems should be considered in the models. In our context, accurate characterizationof factors affecting the plant trophic level (e.g., water) is essential for accurateassessment of the geographic distribution and abundance of higher trophic levels.

Ecological Niche Models based on Physiological Indices (ENPIM)

This approach is based on the concept of plant growth indices proposed by Fitzpatrickand Nix (1968) for Australian pastures. The approach has quasi-mechanistic rootsas it assumes that the normalized growth response of plants to various factors ishump-shaped with minimum and maximum values and an optimum. Hence at anytime t, the overall normalized growth response of a plant (0 ≤ GI(t) < 1) isthe product of indices for the effects of the different resources sought: e.g., light



Fig. 3. Growth index models (cf. Fitzpatrick and Nix, 1968).

(LI(t)), temperature (TI(t)), moisture (MI(t)), nitrogen (NI(t)) and other factors.Low (shortfalls) or high (toxic) concentrations away from the optimum are assumedto limit (sensu von Liebig, 1840) or slow the growth of the species. A hypotheticalexample of several growth indices is illustrated in Fig. 3 where the symbol (©)

indicates the value at time t. In our example, soil moisture is the major limitingfactor, but shortfall or excesses predicted for other factors can also be compoundingcontributing factors Eq. (1).

0 ≤ GI(t) = LI(t) × TI(t) × MI(t) × NI(t) × · · · < 1. (1)

A hypothetical trace of time-varyingTI and MI values over a year at one location isplotted as a dashed line in Fig. 4(a). Superimposed on this trace are the overlappingtolerance limits (i.e., niches) for a plant, an herbivore and a natural enemy thatinteract in a food chain. In our first example (Fig. 4(a)), the time trace of MI andTI is largely within the tolerance limits of favorability of all three species. Whilethe tolerance of a species to a factor changes in evolutionary time, climate maychange more rapidly and cause the trace of the indices to fall outside of the species’tolerance. For example, if the climate in the area becomes hotter and drier, the degreeof overlap may become marginal for the herbivore and the predator while still beingin the suitable range for the plant (Fig. 4(b)). If climate becomes wetter and the rangeof temperature increases (Fig. 4(c)), the trace of the indices could fall largely outsideof the tolerance regions for all three species. Some species may be able to exploit briefperiods of favorability by entering quiescence/dormancy during adverse periods.

The ENPIM approach was used to characterize the limits of favorability forseveral species of aphids in Australian pastures using soil moisture to characterizeplant growth and temperature developmental effects on the aphids, and to maproughly the geographic limits of the cowpea aphid (Aphis craccivora) (Gutierrezet al., 1974; Gutierrez and Yaninek, 1983). Sutherst et al. (1991) embedded thesenotions as an algorithm (CLIMEX) in a geographic information system (GIS) that



Fig. 4. Climate matching of species tolerance limits to temperature (TI) and moisture (MI) at onelocation and the hypothetical trace (- - - -) of index values starting from Julian day 1 (dashed line): (a)good climatic match for all three trophic levels, (b) changes in climatic favorability within the rangelargely favorable for the plant but not for the herbivore or predator, and (c) climate change unfavorablefor all three species.

has been widely used to map the potential geographic distribution of invasive species(e.g., Venette et al., 2000; Coetzee et al., 2009; Rodda et al., 2009; Sutherst andBourne, 2009). Zalucki and van Klinken (2006) provided a recent summary of thebiological and weather variables used in CLIMEX and some applications. Figure 5illustrates the ecological niche boundaries for three species of aphids using observeddaily weather and a mechanistic water balance model (Gutierrez et al., 1974), theincorporation of the ENPIM in a GIS, and the prediction of the potential geographicrange of the specie and its invasive potential into other geographic regions. Theremay be many other physiological dimensions than those of temperature and waterillustrated in the Figs. 4 and 5.

In addition to climate, the distribution and abundance of a species is often limitedby interacting species (Davis et al., 1998; Hodkinson, 1999), but this is difficult toimplement in index models. In addition, weather patterns during short periods of theseason may be critical and affect aspects of behavior, seasonality (e.g., dormancy orquiescence), survival and the dynamics of species in food webs. An example is thebiological control of walnut aphid wherein an introduced strain of parasitoid (Trioxyspallidus) from France failed to control walnut aphid during the hot dry summers in



Fig. 5. Developing an ENM using the growth index approach: (a) ecological niches (ecologicalspace defined using bi-variate normal regression) of three aphids (from Gutierrez et al., 1974); and(b) extrapolating the ecological space to geographic space beyond the original study area (Adaptedfrom Zalucki and van Klinken, 2006).

Central California, while a strain from Iran that was able to enter dormancy providedexcellent control (van den Bosch et al., 1979). More complicated examples includethe cosmopolitan cabbage root fly (Johnsen et al., 1997) and the pink bollworm incotton (Gutierrez et al., 2006a) whose geographical distribution and abundance areseverely limited by temperature and photoperiod that induce dormancy and alter theirphenology and dynamics. Relevant interacting species such as host plants shouldbe included in some analyses to obtain a more realistic estimate of distribution andabundance. However, these factors are difficult to incorporate in ENPIM that oftenuse average weather as drivers.

Medical and veterinary vector-borne diseases

The potential impact of climate change on human diseases is of major concern.Vectorborne diseases imposing large global burdens that are highly sensitive to climate andecological change (Molyneux, 2003) include malaria, schistosomiasis, filariasis,dengue fever, leishmaniasis, Chagas disease, West Nile virus, Lyme disease, tick-borne encephalitis,African trypanosomiasis, onchocerciasis, and cholera (Patz et al.,2005). With some notable exceptions (e.g., tick borne diseases), the effect of climatechange on animal diseases has received comparatively little attention (de la Rocqueet al., 2008). Baylis and Githeco (2006) and Van den Bossche and Coetzer (2008)list the most important environmentally-driven epizootics that are expected to be



influenced by climate change.Analyses of these systems have ranged from statisticalecological models, semi-quantitative methods or simplified dynamical models basedon indexes (e.g., maximum daily reproductive rate of the disease also known as thevectorial capacity) to a few mechanistic models (Patz et al., 2003).

Among vector-borne human diseases, malaria has the greatest potential toincrease its geographic distribution and epidemic patterns in response to climatechange because of the effects on vector ecology and behavior (Martens et al., 1999).Attempts have been made to relate malaria resurgence in East African highlands tochanges in temperature (Pascual et al., 2006) and climate variability caused by the ElNino-Southern Oscillation (ENSO) (Lindblade et al., 1999). However, temperatureis but one part of the factors determining mosquito dynamics (Small et al., 2003),as spatial and temporal variations in rainfall also determine the nature and scale ofmalaria in highland and in low-lying semi-arid areas (Brown et al., 1998; Cox et al.,1999; Kilian et al., 1999; Poveda et al., 2001).

While many interpretations of the major role of climate change on vector-bornediseases are persuasive because they are intuitive, key factors in the transmission andepidemiology of the disease are often ignored (Reiter, 2008), and analyses basedon simple models may not provide a holistic view of the complex ecology anddisease-host interactions. In addition to physical variables, biological, ecologicaland evolutionary processes and dynamics in epidemiological systems need to beconsidered (Smith et al., 2005).

Process-based models offer more opportunities than statistical and semi-quantitative models for realistic representation of both physical and biological inter-acting component in eco-epidemiological systems (Patz et al., 2003). The biologi-cal and ecological realism of physiologically-based demographic models (Gutierrezet al., 1994) would seem to be a promising approach for vector-host-parasite inter-action as the malaria systems (Gilioli et al., 2008; Mariani and Gilioli, 2009).

Physiologically Based Demographic Ecological Niche Models (PBDM )

This approach reverses the process used to develop statistical and physiologicalindex environmental niche models. In these models, the physiological and demo-graphic responses (processes) of the organism (species) to abiotic and biotic fac-tors are modeled on a per capita basis (Gutierrez and Baumgartner, 1984), and themodel(s) at the population level is embedded in a GIS and used to explore thedistribution and abundance of a species. The model driven by appropriate weatherdata and in a tritrophic context is expected to reproduce qualitatively the observedgeographic distribution and abundance of the species.

This approach has its roots in the work of de Wit and Goudriaan (1978) onphysiologically-based plant canopy models. van Ittersum et al. (2003) summarized



de Wit’s concept as follows: defining factors (e.g., CO2, temperature, radiation,and crop characteristics and genetics) determine the potential growth rate, limitingfactors (water and nutrients, etc.) many constrain this potential, and reducing (orbiotic) factors (herbivores, weeds and diseases, pollution) that may further reducegrowth. These factors have parallels in Hutchinson’s concept of the fundamental andrealized niche, but as we shall see PBDM can also include food web interactions. Thephysiologically-based approach assumes that the per capita maximum rate of growthis reduced by conditions the organism experiences and by its interactions with otherspecies. This concept of factors reducing the maximum growth rate to the observedwas proposed independently in entomology by Hughes (1968). Combining the deWit–Hughes notions enables PBDM development at all trophic levels — at the indi-vidual (Gutierrez et al., 1981; 1987), population, food webs and meta-population(Gutierrez and Baumgartner 1984, Gutierrez et al., 1988), and the regional levels(e.g., Gutierrez et al., 2007; 2008; 2009; Ponti et al., 2009; 2010) (Fig. 6(a)).

Tritrophic systems are composed of interacting populations with: (i) the resourceand consumer populations characterized by resource-limited growth; (ii) the plant

Fig. 6. Levels of physiologically based models: (a) individual, population, across ecological zonesand across large geographic regions, and (b) ecosystem analysis using site-specific weather, GIS mapsand marginal analysis.



population integrates the bottom-up effects of weather and edaphic factors (mois-ture, nutrients and others) (Eq. (1)) and is in turn regulated by the top-down effectof consumers that may also be influenced from lower and higher trophic levels(Hairston et al., 1960; Rosenzweig, 1973); (iii) the attack and consumption rate ofall populations are regulated by the same functional response model; and (iv) thelife-history strategies of all populations are forced by environmental-driving vari-ables. Nutritional quality and phenology of the resource population also affect thedemography and phenology of consumer trophic levels.

The same numerical and functional response models are used for all trophic levelsand the linkages between them (Gutierrez and Baumgartner, 1984; Gutierrez, 1992).The underlying mathematics for the population dynamics is found in Gutierrez(1996) and DiCola et al. (1999). The models are driven by observed weather topredict site-specific dynamics or the dynamics across a landscape. When imbeddedin a GIS (e.g., GRASS1), the model can be used to estimate the range and relativeabundance of the species (Fig. 6(b)). Marginal analyses of multi-variate models ofthe simulation data can be used to quantify the trends in the data and may provemore useful than the insights gained from the maps of range shifts alone (Gutierrezet al., 2008).

Food webs in systems as diverse as alfalfa, cassava, cotton, coffee, bean, olive,rice, wheat, yellow starthistle, and aquatic systems have been modeled using thisapproach, and many of the models have been tested with field data and used todevelop IPM strategies and to explore theory (see Gutierrez et al., 2005; Gutierrezand Baumgartner, 2007). Other physiologically-based models include those for for-est defoliators (Williams and Liebhold, 1997; Logan et al., 2007;Abbott and Dwyer,2008), mosquitoes (Focks et al., 1993; Kearney et al., 2009), and bark beetles(Ungerer et al., 1999) and others, but none of these have included the plant level.

Considerable realism can be added in PBDM as illustrated by the examplesbelow and in especially in Gutierrez et al. (2008). PBDMs may be used to explorehow weather and climate change may affect trophic interactions that may deter-mine the geographic range of species. We first examine the effects of temperatureon the control of olive scale (Parlatoria oleae) by two parasitoids. In our sec-ond example, we review the effects of weather and climate change scenarios onthe olive/olive-fly system and the influence their geographic range, phenology andabundance. These examples illustrate some of the applications and processes illus-trated in Fig. 6(b). These unified models are part of a suite called CASAS models(see http://cnr.berkeley.edu/casas).

1Geographic Resources Analysis Support System (GRASS) software was originally developed by the US ArmyCorp of Engineers, but the version we used (GRASS Development Team, 2010; see http://grass.osgeo.org)is maintained and further developed as an official project of the Open Source Geospatial Foundation(http://www.osgeo.org/).



The Olive System

Olive scale/parasitoid system

Weather has considerable influence on the complex developmental biology anddynamics of the sexual olive scale and the two parasitoids that attack it in Californiaolive (Rochat and Gutierrez, 2001; Fig. 7(a)). The asexual parasitoid Aphytismaculicornis produces only females and only on olive scale. In contrast, the sexualparasitoid Coccophagoides utilis produces females on olive scale hosts and malesby parasitizing pupae of female siblings. A. maculicornis operates at lower tempera-tures than C. utilis and is the victor in cases of larval competition within a single host.The model incorporates the effect of temperature on development, reproduction andsurvival of the scale and the parasitoids as well as their preference for olive scalelife stages. The biology and the mathematics of these processes are too extensive toreview here, and hence interested readers are referred to the original article.

Huffaker and Kennett (1966) and Kennett et al. (1966), based on field studiesand intuition, concluded that only the combined action of the two parasitoids couldprovide the high level of control of olive scale observed in Californian olive. Toexplore this, different combinations of maximum and minimum daily temperatureswere used to drive the model to examine the effects on parasitoid regulation andcontrol of olive scale populations — i.e., a demographic ecological niche model wasdeveloped. In the absence of natural enemies, olive scale densities are uniformly highacross all temperature regimes (Fig. 7(b)). However, including the parasitoids one ata time in the model shows that C. utilis was able to suppress olive scale densities onlyin higher max-min temperature regimes (i.e., the hotter parts of the season in the field,Fig. 7(c)), while A. maculicornis alone could suppressed olive scale only in the coolermax-min regimes (i.e., the cooler periods in the field, Fig. 7(d)). The model suggeststhat only the combined action of the two parasitoids acting largely in different butpartially overlapping regions of the max-min temperature state space can contrololive scale (Fig. 7(e)), mimicking the effects of seasonal shifts in temperature thatoccur during the year. The question of whether good biological control of olive scalewill continue in the face of climate change across the large geographic range of thispest requires embedding the model into a geographic information system (GIS).This is demonstrated using the olive-olive fly system.

Olive/olive fly system

Olive and olive fly are of African origin. A review of the biology of both species isfound in Gutierrez et al. (2009), and only a brief overview is presented here.

Olive is a drought-tolerant, long-lived species that exhibits little response to pho-toperiod; its distribution is limited largely by low and high temperatures, and less so



Fig. 7. The biology (a) of the sexual olive scale (OS) and two parasitoids: the sexual autonomoushyperparasitoid (C, Coccophagoides utilis) and asexual thelytokous parasitoid (A, Aphytis maculicor-nis), and (b) the effects of the different combinations of the two parasitoids on olive scale populationdensities plotted on combinations of max-min temperatures (cf. Rochat and Gutierrez, 2001).



by soil water and other factors. Temperature affects the rates of development, pho-tosynthesis, respiration, and subunit initiation rates (i.e., of leaves, stem, root), andmortality. The response of olive to average temperature is non linear, but all plantsubunits are assumed to have the same lower (9.1◦C) and upper (38◦C) thresholds.Note that these thresholds may be exceeded for short periods during the day withmarginal effects. The normalized net of the photosynthetic rate minus the respirationrate over temperature defines the upper and lower thermal thresholds for develop-ment and the optimum. Freezing temperature causes mortality to plant subunits andin the extreme may kill the whole plant.

The maximum number of flower buds in the current season is a function of theamount of fruit wood produced in the previous season, but some of buds may be killedby freezing temperatures. Approximately 450 h < 7.3◦C are required to stimulatespring fruit bud initiation leading to flowering and 400 degree-days (dd) are requiredfrom bud swelling to flowering with the period of flowering being approximately7–10 days. Massive shedding of young post flower fruit occurs due to lack of polli-nation, cold weather, and shortfalls of photosynthate. A mean of 1495 degree-daysare required from flowering to fruit maturation with the fruit becoming susceptibleto olive fly when the seed begins to harden. These parameters vary with variety.

Olive fly is endemic in much of the Mediterranean Basin and Middle East, andwas discovered in 1998 in California, where it is now widely distributed. The biologyof the fly is closely linked to olive fruit age and availability. Adult flies over-winterin facultative reproductive-dormancy that begins to break when fruit of increasinglypreferred ages become available, with preference reaching its maximum when theseed begins to harden. Dormancy is also induced during summer when fruits arein short supply, when mean temperatures fall below 15◦C, and during periods ofhigh summer temperatures. Olive fly females prefer to lay single eggs in unattackedfruit, but multiple attacks may occur when fruit numbers are limiting. When veryripe, fruit fall to the ground and may play an important role in the fly’s dynamics asimmature stages in them continue development during the winter.

The lower temperature thresholds for olive fly were computed from the literature:6.3◦C for egg–larval stages and 8◦C for pupae and adults with the upper thresholdfor all stages being approximately 33◦C. Low and high temperatures increase flymortality in all stages. Four to five generations may develop in highly favorableareas.

The model predicts many aspects of olive and olive-fly dynamics and biology,but only flowering phenology, average total fly pupae per tree, and final infestationrates in the absence of control are reviewed and used to assess the possible effectsof observed weather and three climate change scenarios: Ten years of observedweather in Arizona-California and Italy and three climate change scenarios where



daily weather was increased an average of 1, 2 and 3◦C (see Gutierrez et al. (2009)for full details). The analysis can be done at the field, local area, national or regionalscale.

Olive flowering

Sufficient chilling for flower bud initiation accrued at all locations and years inArizona-California (AZ-CA) and Italy. The range of mean bloom dates using his-torical weather ranged from Julian day 80 to 163, with the earliest bloom dateoccurring in the hotter areas of southern California and Arizona (Fig. 8(a), see his-tograms). In AZ-CA, a dramatic shift toward earlier blooming occurs throughoutmuch of north and central California with increasing temperatures. Specifically, athree-day decrease in bloom dates occurs per 1◦C rise in mean temperature at thelower end of the range, while the effect on the upper end is less clear. Bloomingoccurs later in Italy where the range of bloom dates is 114–178 (Fig. 8(b)). In Italy,the lower and upper end of the range of blooming dates decreases 6–8 days per 1◦C

Fig. 8. Simulated average Julian bloom dates for olive inArizona-California and Italy using observedweather (a, b), and the distribution of bloom dates across both regions using observed weather andthree climate warming scenarios (from Gutierrez et al., 2009).



increase in mean temperature. These changes are seen as shifts in the histogramsfor each increment of temperature.

Olive fly abundance

Olive fly’s dynamics are closely linked to olive fruiting, but as described above thereare independent effects of temperature on development, reproduction, survival anddormancy. Total season-long number of pupae produced is used as a metric ofthe fly’s invasive potential given the bottom-up effects of olive fruit abundance,phenology, and temperature. The infestation levels in California (Fig. 9(a)) coincidequite well with its reported distribution in California (Hannah Nadel and colleagues,http://arcims.uckac.edu/CIMIS/). The areas predicted to be most favorable for the

Fig. 9. Predicted average number of pupae/tree in Arizona–California (a–d) and Italy (e–h)) underobserved and three climate-warming scenarios (+1, +2 and +3◦C) (from Gutierrez et al., 2009). Theright inset is an enlargement for the Garda Lake region in northern Italy (see Fig. 9(e)).



olive fly are along the southern coast of California, where temperatures remainmild throughout the year, while the desert regions of AZ-CA are generally unfavor-able because of high summer temperatures that increase fly mortality and decreasereproduction directly and cause reproductive dormancy. The effects of high tem-perature on olive fly are most obvious in the lower part of the Central Valley ofCalifornia and the deserts of Southern California and Arizona where high summertemperatures allow only low to intermediate densities of the fly to develop. Climatewarming is expected to cause the favorable range for olive fly to contract furtherin the Central Valley and desert regions of AZ-CA, but to increase in coastal areas(Figs. 9(b) to 9(d)).

Linear multiple regression may be used to summarize the GIS data (see Gutierrezet al., 2005; 2008) on olive fly pupae abundance for Arizona–California (Eq. (2)).The coefficients of the regression model for log10 pupae on total ddb < 0◦C, dateof bloom (Blm) and cumulative year long rainfall (mm) were highly significant(p < 0.01), while the significance level for dda was p < 0.05. Only the coefficientfor dda and ddb were negative.

Log10pupae = 5.151 − 0.000646dda − 0.00187ddb + 0.0018Blm + 0.00039mm

df = 6, 567, R2 = 0.44, F = 1288.4

(means: dda = 2612; ddb = 231.4.0; Blm = 104.8, mm = 418.9). (2)

Substituting mean values for the independent variables shows that dda (i.e.,0.000646 × 2612 = −1.688) followed by ddb (−0.433) have the greatest meannegative impact on log pupal density, while Blm (0.189) and mm (0.163) have pos-itive effects resulting in a mean of 2,421.0 pupae per tree across AZ-CA. [As animportant aside, this marginal approach bridges biology and economics as it canbe used to estimate the effects on yield (or other factors) of weather, pests andsay natural enemies in complicated biological systems (see Gutierrez et al., 2005;2006b)].

The restricting effects of cold weather are seen in Northern Italy (Fig. 9(e)). InItaly, only the northern regions and the mountains of central Italy are unfavorable dueto winter temperatures (Fig. 9(e)), but with climate warming, olive fly distributionmoves increasingly northward into previously inhospitable cold areas, and decreasesin the more southern areas due to increased summer temperature (Figs. 9(f ) to 9(h)).The inset in Fig. 9(e) shows the favorable area around Garda Lake in Northern Italywhere olive and olive fly are protected to flourish in what would otherwise be aninhospitable region.

Preliminary results for average olive fly densities and levels of infestations areshown for two decades (e.g., 1956–67 and 1988–97) across the Mediterranean Basin(Fig. 10) (Ponti et al., 2010). The results over this much larger area are more difficult



Fig. 10. Predicted olive fly levels across the Mediterranean Basin (a, b) and the predicted percentageinfested fruit (c, d) during two decades (158–1967 and 1988–1997) (from Ponti et al., 2010).

to summarize and interpret because of the numerous ecological zones that comprisethe region, the grid size of the analysis, and the need to include better water balancemodels for very arid regions. Despite these caveats, the model gives a very reasonablepicture of the distribution of olive fly (Figs. 10(a) and 10(b)). Note that the geographicrange of high infestation levels during the period 1988–97 is greater than for 1956–67(Fig. 10(c) versus Fig. 10(d)).

Lest this section be interpreted as being the final word on appropriate methodol-ogy, we point out some important shortcomings of PBDM: the biology of the relevantinteracting species must be described in a quantitative manner, but often the biolog-ical data to do this are sparse, daily weather data (e.g., max-min temperature, solarradiation, rainfall, RH, wind) are used to drive the model and ideally, calibrated waterand nutrient balance models should be available to capture the bottom-up effects ofthe plant on higher trophic levels. Large scale data to test the model are rarely avail-able. From a mathematical perspective, only numerical solutions of the models arepossible, greatly limiting its use in analysis. However, although development of thesemodels is often deemed difficult and extensive computing capacity to implement themodel is assumed, this is not the case. A comparison with the CLIMEX approach(see Zalucki and van Klinken, 2006) suggests that the number of parameters is less



and that they can all be measured directly with the added benefit that several trophiclevels can be included (see examples in Gutierrez et al., 2008; 2009).

Discussion

Species in all trophic levels will be affected by climate change either directly orindirectly, but the greatest impact is expected to occur when the lower trophic levelsin a food web or chain are affected (i.e., the pyramid of energy). Important changesmay also occur when intermediate levels are adversely affected. The literature isreplete with studies that posit weather effects on species and species interactions,but few of them have been studied in sufficient detail to be useful in analyses of theimpacts of climate change. Because of their economic importance (Pimentel et al.,2000), data on invasive species tend to be more available. For example, weatherwas inferred to influence the dynamics and abundance of the founding case ofmodern biological control in California: the control of cottony cushion scale (Iceryapurchasi, a pest of citrus) by the exotic parasitic fly (Cryptocheatum iceryae) and thepredacious vedalia beetle (Rodalia cardinalis) (Quesada and DeBach, 1973). Thedistribution of the fly C. iceryae is restricted to the cooler near-coastal areas, whilethe vedalia beetle is active across a wide range of ecosystems, including the hotterinland areas of California. However, to determine the geographic ranges of thesespecies and estimate how climate change might affect them requires more biologicalinformation than is currently in the literature. The displacement of successive speciesof red scale parasitoids by later introductions that were better adapted to conditionsof Southern California (Rosen and DeBach, 1979) was analyzed by Murdoch et al.(1996), but the analysis was not extended to the regional level.

Surprisingly, the effects of weather on host plants and the cascading-up effectson higher trophic interactions are usually not included in ENM studies. The classicexample is the outbreak of locust when the confluence of rains increases host plantabundance in space and time, and favorable winds occur that carry building locustpopulations from one favorable place to another until massive swarms occur (Roffeyand Popov, 1968). Rapid buildups in response to host plant availability may occur insome species (e.g., cowpea aphid; Gutierrez et al. (1974); and African armyworms,see Janssen (1993)), while in other cases outbreaks develop more slowly due tocyclic changes in host plants over time [e.g., spruce budworm (Thompson et al.,1979); bark beetles (Rebetez and Dobbertin, 2004); some vertebrate species (e.g.,snowshoe hare/lynx, Krebs et al., 2002) and others]. The capacity to model suchsystems realistically is critical if we are to assess the impact of climate change onnative and invasive species, but these require sound models of the plant dynamics.



RS satellite data hold unprecedented potential for assessing the effects of cli-mate change on ecosystems on regional and global scales, but they need integrationwith reliable field data (Kerr and Ostrovsky, 2003) in an interdisciplinary framework(Muraoka and Koizumi, 2009). Among the most widely used RS data are measuresof canopy greenness called NDVI (Normalized Difference Vegetation Index). Suc-cessful applications of NDVI were reviewed showing the potential of this approachto the study of climate change effects on species in managed and natural systems.

Ecological niche/climate envelope modeling approaches are used to examine theeffects of weather variables on the geographic distribution and abundance of speciesin a food web context. These approaches may be statistical or physiological indexbased, but in both cases the area of current distribution of a species is characterizedclimatically using observed weather or climate model projections. The models maybe used to project the potential range of the species and its invasive potential intoother areas (sensu Estrada-Pena, 2008; Mitikka et al., 2008). Some underpinningassumptions of statistical ENMs include: the current distribution of a species is thebest indicator of climatic requirements; the distributions are in equilibrium withcurrent climate (Guisan and Zimmermann, 2000; Guisan and Thuiller, 2005); andclimate niche conservatism is maintained in both space and time (Peterson et al.,1999). These may hold true for some species, but not for other. Feedbacks of otherinteracting species may help determine the geographic range of a species, but thisis difficult to include in ENMs. Araujo and Luoto (2007) provide support that bioticinteractions significantly affect both the explanatory and predictive powers of ENMmodels at macro scales.

A variety of physiologically-based ENMs have been used to predict the potentialdistribution of invasive species (Coetzee et al., 2009; Rodda et al., 2009; Sutherst andBourne, 2009). These physiological based models seek to describe the physiologicalresponse of plant and animal species to weather and abiotic variables as growthindices. These indices may be developed by fitting hump-shaped response curvesfor species based on their known distribution or they could be determined usinglaboratory or field studies (Gutierrez et al., 1974). The best known of these systemsis the widely used CLIMEX system (Sutherst et al., 1991). Including feedback ofinteracting species that are determinants of the geographic range of a species isdifficult (Davis et al., 1998).

There have been many physiologically-based demographic models of localdynamics that include biotic interactions (e.g., Gutierrez and Baumgartner, 1984;Wu and Wilson, 1997; Holst and Ruggle, 1997) and that can be extended to regionalanalyses (e.g., Gutierrez et al., 2005; 2006a,b; 2007; 2009; Ponti et al., 2009). Theability to include biotic interactions and feedback potentially makes them moreaccurate for assessing the geographic distribution and abundance of species, andfor examining the potential effects of climate change on them. Drawbacks to this



approach include the need for more extensive biological data to formulate the modeland more weather data at closer intervals to run the model.

When available, these physiologically-based demographic models often givevery different predictions than the physiological index approach. For example,Venette et al. (2000) using CLIMEX concluded that abiotic factors did not pre-clude the establishment of the pink bollworm in cotton over much of the cottongrowing regions of the SE United States, and that its absence in “favorable regions”may be the result of federal monitoring, quarantine, and eradication programs butdid not rule out climate and other ecological factors. In contrast, using PBDM thatincluded the dynamics of cotton and the factors affecting PBW reproduction, devel-opment, survival, and induction and termination of diapause gave a more restricteddistribution for this pest in Arizona and California. This analysis also predicted thata 2.5◦C increase in temperatures would increase considerably the range of this pestin California.

The newly-discovered polyphagous light brown apple moth (LBAM) (Epiphyaspostvittana) in the area around the San Francisco Bay prompted a quarantine ofthe infested area and the initiation of an eradication program. Using two simplecriteria — degree days sufficient for the development of three or more LBAMgenerations and winter temperatures that do not drop below −16◦C for one full dayto ensure winter survival — Fowler et al. (2009) predicted that most of Californiaand the southern half of the United States could be invaded by this pest. In contrast,Gutierrez, Mills and Ponti (2010), using biological data from Danthanarayana (1975)in a PBDM, showed that the distribution of LBAM in California was restrictedlargely to near-coastal areas with some extension eastward into the northern part ofthe Central Valley. Most of the Central Valley and southern Arizona are of less thanintermediate favourability, particularly when high summer temperatures exceed thetolerance of LBAM. Including the limiting effect of rainfall on annual host plants(e.g., Gutierrez et al., 2005) further restricts the distribution of the pest to the coastalregions. Climate warming would further restrict LBAM’s distribution in centralCalifornia, but could extend it northward along the coast.

Conclusion

While the goal of all of the approaches used is to analyze climate change effectson the distribution and possibly abundance of species (invasive and otherwise), themethods cannot be applied with equal success. Each of the methods has differentdata requirements and will provide different insights into how climate change mayaffect host-pest interactions. Hence the method of analysis selected will dependon data availability and on the technical expertise and bias of the researcher(s).



On the positive side, many problems may be solved using more direct or simplerapproaches such as remote sensing or ecological niche models obviating the needto include more biological detail. However, for analyses of tri-trophic systems,the details may matter. Considerable care must be exercised in modeling approachselection, construction and evaluation because the model will always give an answerbut it may simply be wrong (sensu Wang and Gutierrez, 1980).

References

Abbott, K.C., and G. Dwyer. 2008. Using mechanistic models to understand synchrony in forest insectpopulations: the North American gypsy moth as a case study. Amer. Nat. 172:613–624.

Andrewartha H.G., and L.C. Birch. 1954. The Distribution and Abundance of Animals. University ofChicago Press, Chicago.

Anyamba, A., J.P. Chretien, J. Small, C.J. Tucker, P.B. Formenty, J.H. Richardson, S.C. Britch, D.C.Schnabel, R.L. Erickson, and K.J. Linthicum. 2009. Prediction of a Rift Valley fever outbreak.PNAS, 106:955–959.

Araujo, M.B., and M. Luoto. 2007. The importance of biotic interactions for modelling speciesdistributions under climate change. Global Ecol. Biogeogr. 16:743–753.

Araujo, M.B., and A. Guisan. 2006. Five (or so) challenges for species distribution modelling. J. Bio-geogr. 33:1677–1688.

Battisti,A., M. Stastny, S. Netherer, C. Robinet,A. Schopf,A. Roques, and S. Larsson. 2005. Expansionof geographic range in the pine processionary moth caused by increased winter temperatures.Ecol. Appl. 15:2084–2096.

Baylis, M., and A.K. Githeko. 2006. The effects of climate change on infectious diseases in animals.Report T7.3 for the UK Government’s Foresight project, Infectious Diseases: preparing for thefuture. Department of Trade and Industry, UK, 1–35.

Beaumont, L.J., R.V. Gallagher, W. Thuiller, P.O. Downey, M.R. Leishman, and L. Hughes. 2009.Different climatic envelopes among invasive populations may lead to underestimations ofcurrent and future biological invasions. Diversity Distrib. 15:409–420.

Bradley, B.A., and E. Fleishman. 2008. Can remote sensing of land cover improve species distributionmodelling? J. Biogeogr. 35:1158–1159.

Brown, V.M., I. Abdir, M. Rossi, P. Barboza, and A. Paugam. 1998. Epidemic of malaria in north-eastern Kenya. Lancet 352:1356–1357.

Camps-Valls, G., J. Munoz-Marı, L. Gomez-Chova, K. Richter, and J. Calpe-Maravilla. 2009.Biophysical parameter estimation with a semisupervised support vector machine. IEEE Geosci.Remote Sensing Letters 6:248–252.

Carpi, G., F. Cagnacci, M. Neteler, and A. Rizzoli. 2007. Tick infestation on roe deer in relation togeographic and remotely sensed climatic variables in a tick-borne encephalitis endemic area.Epidemiology & Infection 136:1416–1424.

Cherlet, M.R.,A. Di Gregorio, and J.U. Hielkema. 1991. Remote-sensing applications for desert-locustmonitoring and forecasting. EPPO Bull. 21:633–642.

Clark, J.S., S.R. Carpenter, M. Barber, S. Collins, A. Dobson, J.A. Foley, D.M. Lodge, M. Pascual,R. Pielke Jr., W. Pizer, C. Pringle, W.V. Reid, K.A. Rose, O. Sala, W.H. Schlesinger, D.H. Wall,and D. Wear. 2001. Ecological forecasts: an emerging imperative. Science, 293:657–660.

Coetzee, J.A., M.P. Hill, and D. Schlange. 2009. Potential spread of the invasive plant Hydrilla verti-cillata in South Africa based on anthropogenic spread and climate suitability. Biol. Invasions11:801–812.



Coops, N.C., R.H. Waring, M.A. Wulder, and J.C. White. 2009. Prediction and assessment of barkbeetle-induced mortality of lodgepole pine using estimates of stand vigor derived from remotelysensed data. Remote Sensing Environ. 113:1058–1066.

Cox J., M.H. Craig, D. le Sueur, and B. Sharp. 1999. Mapping malaria risk in the highlands of Africa.Mapping Malaria Risk inAfrica/Highland Malaria Project (MARA/HIMAL). Technical Report,MARA/Durban, London, London School of Hygiene and Tropical Medicine.

Cressman, K. 1996. Current methods of desert locust forecasting at FAO. EPPO Bull. 26:577–585.Danthanarayana, W. 1975. The bionomics, distribution and host range of the light brown apple moth,

Epiphyas postvittana (Walk.) (Tortricidae), Aust. J. Zool. 23:419–437.Davis, A.J., L.S. Jenkinson, J.H. Lawton, B. Shorrocks, and S. Wood. 1998. Making mistakes when

predicting shifts in species range in response to global warming. Nature 391:783–786.de La Rocque, S., J.A. Rioux, and J. Slingenbergh. 2008. Climate change: effect on animal disease

systems and implications for surveillance and control. Revue Scientifique et Tecnique de l’OfficeInternational des Epizooties 27(2):339–354.

deWit, C.T., and J. Goudriaan. 1978. Simulation of Ecological Processes, 2nd edn. PUDOC Publishers,Wageningen.

DeFries, R. 2008. Terrestrial vegetation in the coupled human-earth system: contributions of remotesensing. Ann. Rev. Environ. Resour. 33:369–390.

DiCola, G., G. Gilioli, and J. Baumgartner. 1999. Mathematical models for age-structured populationdynamics, pp. 503–531. In C.B. Huffaker, and A.P. Gutierrez (eds.), Ecological Entomology,second edition, John Wiley and Sons, New York.

Eklundh, L., T. Johansson, and S. Solberg. 2009. Mapping insect defoliation in Scots pine with MODIStime-series data. Remote Sensing Environ. 113:1566–1573.

Estrada-Pena, A. 2008. Climate, niche, ticks, and models: what they are and how we should interpretthem. Parasitol. Res. 103:87–95.

Fitzpatrick, E.A., and H.A. Nix. 1968. The climatic factor in Australian grasslands ecology, pp. 3–26.In R.M. Moore (ed.), Australian Grasslands. Australian National University Press.

Focks, D.A., D.G. Haile, E. Daniels, G.A. Mount. 1993. Dynamic life table model for Aedes aegypti(Diptera: Culicidae): Analysis of the literature and model development. J. Med. Entomol.30:1003–1017.

Ford, T.E., R.R. Colwell, J.B Rose, S.S. Morse, D.J. Rogers, and T.L. Yates. 2009. Using satelliteimages of environmental changes to predict infectious disease outbreaks. Emerg. Infect. Dis.doi: 10.3201/eid1509.081334.

Fowler, G., L. Garrett, A. Neeley, R. Margarey, D. Borchert, and B. Spears. 2009. Economic analysis:risk to U.S. apple, grape, orange and pear production from the light brown apple moth, Epiphyaspostvittana (Walker). USDA-APHIS-PPQ-CPHST-PERAL, pp. 28.

Galan, C., H. Garcia-Mozo, L. Vazquez, L. Ruiz, C.D. de la Guardia, and M.M. Trigo. 2005. Heatrequirement for the onset of the Olea europaea L. pollen season in several sites in Andalusiaand the effect of the expected future climate change. Internat. J. Biomet. 49:184–188.

Gilioli, G., J. Baumgartner, G. DiCola,A.P. Gutierrez, H. Herren, B. Lindtjørn, L. Mariani, S. Pasquali,and D. Wakgari. 2008. Across spatial scales dynamics of Anopheles gambiae populations andmalaria transmission. Proc. XXIII Internat. Congr. Entomol., Durban, South Africa, 6–12 July,2008, p. 772.

Gordo, O., and J.J. Sanz. 2005. Phenology and climate change: a long-term study in a Mediterraneanlocality. Oecologia 146:484–495.

Gordo, O., and J.J. Sanz. 2009. Long-term temporal changes of plant phenology in the WesternMediterranean. Global Change Biol. 15:1930–1948.

Guisan, A., and N.E. Zimmermann. 2000. Predictive habitat distribution models in ecology. Ecol.Modelling 135:147–186.

Guisan, A., and W. Thuiller. 2005. Predicting species distribution: offering more than simple habitatmodels. Ecol. Letters 8:993–1009.



Gutierrez, A.P. 1992. The physiological basis of ratio dependent theory. Ecology 73:1552–1563.Gutierrez, A.P. 1996. Applied Population Ecology: A Supply-Demand Approach. John Wiley and

Sons, New York.Gutierrez, A.P., and J.U. Baumgartner. 1984. Multi-trophic level models of predator-prey-energetics:

I. Age specific energetics models-pea aphid Acyrthosiphon pisum (Harris) (Homoptera: Aphi-didae) as an example. Can. Entomol. 116:924–932.

Gutierrez,A.P., and J. Baumgartner. 2007. Modeling the dynamics of tritrophic population interactions,pp. 301–360. In M. Kogan and P. Jepson (eds.), Perspectives in Ecology and Integrated PestManagement, Cambridge University Press, Cambridge.

Gutierrez, A. P., J.U. Baumgartner, and K. S. Hagen. 1981. A conceptual model for growth, develop-ment, and reproduction in the ladybird beetle, Hippodamia convergens (Coleoptera: Coccinel-lidae). Can. Entomol. 113:21–33.

Gutierrez, A.P., K.M. Daane, L. Ponti, V.M. Walton, and C.K. Ellis. 2008. Prospective evaluation ofthe biological control of vine mealybug: refuge effects and climate. J. Appl. Ecol. 45:524–536.

Gutierrez, A.P., C.K. Ellis, T. d’Oultremont, and L. Ponti. 2006a. Climatic limits of pink bollworm inArizona and California: effects of climate warming. Acta Oecologica 30:353–364.

Gutierrez, A.P., P. Neuenschwander, F. Schulthess, H.R. Herren, J.U. Baumgartner, B. Wermelinger,J.S. Yaninek, and C.K. Ellis. 1988. Analysis of biological control of cassava pests in Africa: II.Cassava mealybug Phenacoccus manihoti. J. Appl. Ecol. 25:921–940.

Gutierrez, A.P., N.J. Mills, and L. Ponti (2010). Limits to the potential distribution of light brown applemoth in Arizona-California based on climate suitability and host plant availability. BiologicalInvasions, doi 10.1007/s10530-010-9725-8.

Gutierrez, A. P., N. J. Mills, S. J. Schreiber, and C. K. Ellis. 1994. A physiologically based tritrophicperspective on bottom up–top down regulation of populations. Ecology 75:2227–2242.

Gutierrez, A.P., H.A. Nix, D.E. Havenstein, and P.A. Moore. 1974. The ecology of Aphis craccivoraKoch and Subterranean Clover Stunt Virus in south-east Australia. III. A regional perspectiveof the phenology and migration of the cowpea aphid. J. Appl. Ecol. 11:21–35.

Gutierrez,A.P., M.J. Pitcairn, C.K. Ellis, N. Carruthers, and R. Ghezelbash. 2005. Evaluating biologicalcontrol of yellow starthistle (Centaurea solstitialis) in California: A GIS based supply–demanddemographic model. Biol. Control 34: 115–131.

Gutierrez, A.P., L. Ponti, and Q.A. Cossu. 2009. Effects of climate warming on olive and olive fly.(Bactrocera oleae (Gmelin)) in California and Italy. Climatic Change 95:195–217.

Gutierrez, A.P., L. Ponti, C.K. Ellis, and T. d’Oultremont. 2006b. Analysis of climate effects onagricultural systems: A report to the Governor of California sponsored by the California Cli-mate Change Center. http://www.energy.ca.gov/2005publications/CEC-500-2005-188/CEC-500-2005-188-SF.pdf.

Gutierrez, A.P., L. Ponti, T. d’Oultremont, T. d’Oultremont, and C.K. Ellis. 2008. Climate changeeffects on poikilotherm tritrophic interactions. Climatic Change 87:167–192.

Gutierrez, A.P., F. Schulthess, L.T. Wilson, A.M. Villacorta, C.K. Ellis, and J.U. Baumgartner. 1987.Energy acquisition and allocation in plants and insects: A hypothesis for the possible role ofhormones in insect feeding patterns. Can. Entomol. 119:109–129.

Gutierrez, A.P., and J.S.Yaninek. 1983. Responses to weather of eight aphid species commonly foundin pastures in southeastern Australia. Can. Entomol. 115:1359–1364.

Hairston N.G., F.E. Smith, and L.B. Slobodkin. 1960. Community structure, population control, andcompetition. Am. Nat. 94:421–5.

Hickler, T., S. Fronzek, M.B. Araujo, O. Schweiger, W. Thuiller, and M.T. Sykes. 2009. An eco-system model-based estimate of changes in water availability differs from water proxies thatare commonly used in species distribution models. Global Ecol. Biogeogr. 18:304–313.

Hielkema, J.U., J. Roffey, and C.J. Tucker. 1986. Assessment of ecological conditions associated withthe 1980/81 desert locust plague upsurge in West Africa using environmental satellite data.Internat. J. Remote Sensing 7:1609–1622.



Hodkinson, I.D. 1999. Species response to global environmental change or why ecophysiologicalmodels are important: a reply to Davis et al. J. Anim. Ecol. 68:1259–1262.

Holling, C.S. 1992. Cross-scale morphology, geometry, and dynamics of ecosystems. Ecol. Mono-graphs 62:447–502.

Holst, N., and P. Ruggle. 1997. A physiologically based model of pest–natural enemy interactions.Exp. Appl. Acarol. 21:325–41.

Huffaker, C.B., and C.E. Kennett. 1966. Studies of two parasites in control of the olive scale, Parlatoriaoleae (Colvee) IV. Biological control of Parlatoria oleae (Colvee) through the compensatoryaction of two introduced parasites. Hilgardia 37:283–334.

Hughes, R.D. 1963. Population dynamics of the cabbage aphid. J. Anim. Ecol. 32:393–424.Hutchinson, G.E. 1957. Concluding remarks. Cold Spring Harbor Symposia on Quantitative Biology

22:415–27.Janssen J.A.M. 1993. Soil nutrient availability in a primary outbreak area of the African armyworm,

Spodoptera exempta (Lepidoptera, Noctuidae), in relation to drought intensity and outbreakdevelopment in Kenya. Bull. Entomol. Res. 83:579–593.

Jepsen, J.U., S.B. Hagen, K.A. Høgda, R.A. Ims, S.R. Karlsen, H. Tømmervik, and N.G. Yoccoz.2009. Monitoring the spatio-temporal dynamics of geometrid moth outbreaks in birch forestusing MODIS-NDVI data. Remote Sensing Environ. 113:1939–1947.

Ji, R., B.Y. Xie, D.M. Li, Z. Li, and X. Zhang. 2004. Use of MODIS data to monitor the orientalmigratory locust plague. Agric. Ecosys. Environ. 104:615–620.

Johnsen, S., A.P. Gutierrez, and J. Jorgensen. 1997. Overwintering in the cabbage root fly Deliaradicum: A dynamic model of temperature-dependent dormancy and post-dormancy develop-ment. J. Appl. Ecol. 34:21–28.

Kalluri, S., P. Gilruth, D. Rogers, and M. Szczur. 2007. Surveillance of arthropod vector-borne infec-tious diseases using remote sensing techniques: a review. PLoS Pathog. 3:116.

Kearney, M., W.P. Porter, C. Williams, S. Ritchie, and A.A. Hoffmann. 2009. Integrating biophysicalmodels and evolutionary theory to predict climatic impacts on species’ ranges: the denguemosquito Aedes aegypti in Australia. Funct. Ecol. 23: 528–538.

Kennett, C.E., C.B. Huffaker, and G.L. Finney. 1966. Studies of two parasites of the olive scale,Parlatoria oleae (Colvee) III. The role of an autoparasitic aphelinid, Coccophagoides utilisDoutt, in the control of the olive scale, Parlatoria oleae (Colvee). Hilgardia 37:255–282.

Kerr, J.T., and M. Ostrovsky. 2003. From space to species: ecological applications for remote sensing.TREE 18:299–305.

Kilian A.H., P. Langi, A. Talisuna, and G. Kabagambe. 1999. Rainfall pattern, El Nino and malaria inUganda. Trans. R. Soc. Trop. Med. Hyg. 93:22–23.

Krebs C.J., T.N. Zimmerling, C. Jardine, K.A. Trostel, A.J. Kenney, S. Gilbert, and E.J. Hofer. 2002.Cyclic dynamics of snowshoe hares on a small island in theYukon. Can. J. Zool. 80:1442–1450.

Larcher, W. 1995. Physiological Plant Ecology, 3rd edn. Berlin, Springer.Latimer, A.M., S. Wu, A. Gelfand, and J.A. Silander Jr. 2006. Building statistical models to analyze

species distributions. Ecol. Appl. 16:33–50.Liang, L., and M.D. Schwartz. 2009. Landscape phenology: an integrative approach to seasonal

vegetation dynamics. Landscape Ecol. 24:465–472.Lindblade K.A., E.D. Walker, A.W. Onapa, J. Katungu, and M.L. Wilson. 1999. Highland malaria

in Uganda: prospective analysis of an epidemic associated with El Nino. Trans. R. Soc. Trop.Med. Hyg. 93:480–487.

Linthicum, K.J., C.L. Bailey, F.G. Davies, and C.J. Tucker. 1987. Detection of Rift Valley fever viralactivity in Kenya by satellite remote sensing imagery. Science 235:1656–1659.

Liu, Z., X. Shi, E. Warner, Y. Ge, D. Yu, S. Ni, and H. Wang. 2008. Relationship between orientalmigratory locust plague and soil moisture extracted from MODIS data. Intern. J. Appl. EarthObserv. Geoinf. 10:84–91.



Logan J.A., J. Regniere, D.R. Gray, and A.S. Munson. 2007. Risk assessment in face of a changingenvironment: gypsy moth and climate change in Utah. Ecol Appl. 17:101–117.

Ma, J., X. Han, C. Wang, Y. Zhang, J. Tang, Z. Xie, and T. Deveson. 2005. Monitoring East Asianmigratory locust plagues using remote sensing data and field investigations. Intern. J. RemoteSensing 26:629–634.

Mariani, L., and G. Gilioli (submitted). An eco-epidemiological approach to malaria: the meteo-hydrological modules. Malaria J.

Martens P., R.S. Kovats, S. Nijhof, P. de Vries, M.T.J. Livermore, D.J. Bradley, J. Cox, and A.J.McMichael. 1999. Climate change and future populations at risk of malaria. Global Environ.Change 9:89–107.

McCulloch, L., and D.M. Hunter. 1983. Identification and monitoring of Australian plague locusthabitats from Landsat. Remote Sensing Environ. 13:95–102.

Mendelsohn, R., P. Kurukulasuriya, A. Basist, F. Kogan, and C. Williams. 2007. Climate analysis withsatellite versus weather station data. Climatic Change 81:71–83.

Mitikka, V., R.K. Heikkinen. M. Luoto, M.B. Araujo, K. Saarinen, J. Poyry, and S. Fronzek. 2008.Predicting range expansion of the map butterfly in Northern Europe using bioclimatic models.Biodivers. Conserv. 17:623–641.

Molyneux D.H. 2003. Climate change and tropical disease. Trans. R. Soc. Trop. Hyg. 97:129–132.Muraoka, H., and H. Koizumi. 2009. Satellite Ecology (SATECO) — linking ecology, remote sensing

and micrometeorology, from plot to regional scale, for the study of ecosystem structure andfunction. J. Plant Res. 122:3–20.

Murdoch, W.W., C.J Briggs, and R.M. Nisbet. 1996. Competitive displacement and biological controlin parasitoids: a model. Am. Nat. 148:807–826.

Nadel, H., M.W. Johnson, and K. Lynn-Patterson. 2009. Using climate maps in olive fly managementdecisions, http://arcims.uckac.edu/CIMIS/.

Pascual M., J.A. Ahumada, L.F. Chaves, X. Rodo, and M. Bouma. 2006. Malaria resurgence in theEast African highlands: Temperature trends revisited. PNAS 103:5829–5834.

Patz, J.A., U.E.C. Confalonieri, F.P. Amerasinghe, K.B. Chua, P. Daszak, A.D. Hyatt, D. Molyneux,M.Thomson, L.Yameogo, M.M. Lazaro, P.Vasconcelos,Y. Rubio-Palis, D. Campbell-Lendrum,T. Jaenisch, H. Mahamat, C. Mutero, D.Waltner-Toews, and C. Whiteman. 2005. Human health:ecosystem regulation of infectious diseases. in Ecosystems and Human Well-Being: CurrentState and Trends. Volume 1, Washington DC, The Millennium Ecosystem Assessment, IslandPress, 391–415 pp.

Patz J.A., A.K. Githeko, J.P. McCarty, S. Hussein, and U. Confalonieri. 2003. Climate change andinfectious diseases, in Climate Change and Human Health: Risks and Responses. McMichael,A. et al. (eds.), Geneva, WHO, 103–132 pp.

Pengra, B.W., C.A. Johnston, and T.R. Loveland. 2007. Mapping an invasive plant, Phragmites aus-tralis, in coastal wetlands using the EO-1 Hyperion hyperspectral sensor. Remote Sensing ofEnviron. 108:74–81.

Peterson, A.T., J. Soberon, and V. Sanchez-Cordero. 1999. Conservatism of ecological niches inevolutionary time. Science 285:1265–1267.

Pettorelli, N., J.O. Vik, A. Mysterud, J.M. Gaillard, C.J. Tucker, and N.C. Stenseth. 2005. Usingthe satellite-derived NDVI to assess ecological responses to environmental change. TREE 20:503–510.

Pimentel, D., L. Lach, R. Zuniga, and D. Morrison. 2000. Environmental and economic costs ofnonindigenous species in the United States. BioScience 50:53–65.

Ponti, L.,A.P. Gutierrez, and P.M. Ruti. 2010. The olive-Bactrocera oleae (Diptera Tephritidae) systemin the Mediterranean Basin: a physiologically based analysis driven by the ERA-40 climate data.Notiziario sulla Protezione delle Piante (in press).



Ponti, L., Q.A. Cossu, and A.P. Gutierrez. 2009. Climate warming effects on the olive-Bactroceraoleae system on Mediterranean Islands: Sardinia as an example. Global Climate Change.http://dx.doi.org/10.1111/j.1365-2486.2009.01938.x

Poveda G., W. Rojas, M.L. Quinones, I.D. Velez, R.I. Mantilla, D. Ruiz, S.J. Zuluaga, and G.L. Rua.2001. Coupling between annual and ENSO timescales in the malaria-climate association inColombia. Environ. Health Perspectives 109:489–493.

Quezada, J.R., and P. DeBach. 1973. Bioecological and population studies of the cottony scale, Iceryapurchasi Mask. and its natural enemies. Rodolia cardinalis Mul. and Cryptochaetum iceryaeWil., in southern California. Hilgardia 41:631–688.

Rebetez, M., and M. Dobbertin. 2004. Climate change may already threaten Scots pine stands in theSwiss Alps. Theor. Appl. Climatol. 79:1–9.

Reiter P. 2008. Global warming and malaria: knowing the horse before hitching the cart. Malaria J.7:1–9.

Rizzoli, A., M. Neteler, R. Rosa, R., W. Versini, A. Cristofolini, M. Bregoli, A. Buckley, and E.A.Gould. 2007. Early detection of tick-borne encephalitis virus spatial distribution and activityin the province of Trento, northern Italy. Geospatial Health 2:169–176.

Rochat, J., and A.P. Gutierrez. 2001. Weather mediated regulation of olive scale by two parasitoids.J. Anim. Ecol. 70:476–490.

Rodda, G.H., C.S. Jarnevich, and R.N. Reed 2009 What parts of the US mainland are climaticallysuitable for invasive alien pythons spreading from Everglades National Park? Biol. Invasions11:241–252.

Roffey, J., and G. Popov. 1968. Environmental and behavioural processes in desert locust outbreaks.Nature 219:446–450.

Rogers, D.J., S.E. Randolph, R.W. Snow, and S.I. Hay. 2002. Satellite imagery in the study and forecastof malaria. Nature 415:710–715.

Rosen, D., and P. DeBach. 1979. Species of Aphytis of the world (Hymenoptera: Aphelinidae). Junk,Dordrecht.

Rosenzweig, M.L. 1973. Exploitation in three trophic levels. Am. Nat. 107:275–94.Roura-Pascual, N., L. Brotons, A. Townsend Peterson, and W. Thuiller. 2009. Consensual predictions

of potential distributional areas for invasive species: a case study ofArgentine ants in the IberianPeninsula. Biol. Invasions 11:1017–1031.

Schreiber, S.J., and A.P. Gutierrez. 1998. A supply–demand perspective of species invasions andcoexistence: applications to biological control. Ecol. Modelling 106:27–45.

Shippert, M.M., D.A. Walker, N.A. Auerbach, and B.E. Lewis. 1995. Biomass and leaf-area indexmaps derived from SPOT images for Toolik Lake and Imnavait Creek areas, Alaska. PolarRecord 31:147–154.

Sivanpillai, R., and A. Latchininsky. 2007. Mapping locust habitats in the Amudarya River Delta,Uzbekistan with multi-temporal MODIS Imagery. Environ. Management 39:876–886.

Small J., S.J. Goetz, and S.I. Hay. 2003. Climatic suitability for malaria transmission in Africa, 1911–1995. PNAS 100:15341–15345.

Smith K.F., A.P. Dobson, F.E. McKenzie, L.A. Real, D.L. Smith, and M.L. Wilson. 2005. Ecologicaltheory to enhance infectious disease control and public health policy. Frontiers Ecol. Environ.3(1):29–37.

Sutherst, R.W., G.F. Maywald, and W. Bottomly. 1991. From CLIMEX to PESKY, a generic expertsystem for risk assessment. EPPO Bull. 21:595–608.

Sutherst, R.W., and A.S. Bourne. 2009. Modelling non-equilibrium distributions of invasive species:a tale of two modelling paradigms. Biol. Invasions 11: 1231–1237.

Thompson W.A., C.S. Holling, D. Kira, C.C. Huang, and I. Vertinsky. 1979. Evaluation of Alterna-tive Forest System Management Policies — Case of the Spruce budworm in New-Brunswick.J. Environ. Econ. Management 6:51–68.



Thuiller W., D.M. Richardson, P. Pysek, G.F. Midgley, G.O. Hughes, and M. Rouget. 2005. GlobalChange Biol. 11:2234–2250.

Tougou, D., D.L. Musolin, and K. Fujisaki. 2009. Some like it hot! Rapid climate change promotesarea changes in distribution ranges of Nezara viridula and Nezara antennata in Japan, Entomol.Exp. et Appl. 130:249–258.

Tucker, C.J., J.E. Pinzon, M.E. Brown, D.A. Slayback, E.W. Pak, R. Mahoney, E.F. Vermote, and N.E.Saleous. 2005. An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOTvegetation NDVI data. Internat. J. Remote Sensing 26:4485–4498.

Turner, W., S. Spector, N. Gardiner, M. Fladeland, E. Sterling, and M. Steininger. 2003. Remotesensing for biodiversity science and conservation. TREE 18:306–314.

Ungerer, M.J., M.P. Ayres, and M.J. Lombardero. 1999. Climate and the northern distribution limitsof Dendroctonus frontalis Zimmermann (Coleoptera: Scolytidae). J. Biogeogr. 26:1133–1145.

van den Bosch, R., R. Hom, P. Matteson, B.D. Frazer, P.S. Messenger, and C.S. Davis. 1979. Biologicalcontrol of the walnut aphid in California: impact of the parasite, Trioxys pallidus. Hilgardia47:1–3.

Van den Bossche, P., and J.A.V. Coetzer. 2008. Climate change and animal health in Africa. RevueScientifique et Tecnique de l Office International des Epizooties 27(2):551–562.

van Ittersum, M.K., P.A. Leffelaar, H. van Keulen, M.J. Kropff, L. Bastiaans, and J. Goudriaan. 2003.On approaches and applications of the Wageningen crop models. Euro. J. Agron. 18:201–234.

Venette, R.C., S.E. Naranjo, and W.D. Hutchison. 2000. Implications of larval mortality at low temper-atures and high soil moistures for establishment of pink bollworm (Lepidoptera: Gelechiidae)in Southeastern United States cotton. Environ. Entomol. 29:1018–1026.

von Liebig, J. 1840. Chemistry and its Applications to Agriculture and Physiology (4th edn 1847)Taylor and Walton, London.

Voss, F., and U. Dreiser. 1997. Mapping of desert locust habitats using remote sensing techniques.New Strategies in Locust Control (eds. S. Krall, R. Peveling, and D.B. Diallo), pp. 37–45.Birkhauser, Basel, Switzerland.

Walther, G.R. 2002. Ecological responses to recent climate change. Nature 416:389–395.Wang,Y., andA.P. Gutierrez. 1980.An assessment of the use of stability analysis in population ecology.

J. Anim. Ecol. 49:435–452.Wellington, W.G., D.L. Johnson, and D.J. Lactin. 1999. Weather and insects, pp. 313–353. In C.B. Huf-

faker, and A. P. Gutierrez (eds.), Ecological Entomology, second edition. John Wiley and Sons,New York.

Williams, D.W., and A.M. Liebhold. 1997. Latitudinal shifts in spruce budworm (Lepidoptera: Tor-tricidae) outbreaks and spruce-fir forest distributions with climatic change. Acta Phytopathol.Entomol. Hungarica, 32:205–215.

Wolfe D.W., M.D. Schwartz, A.N. Lakso, Y. Otsuki, R.M. Pool, and N. J. Shaulis. 2005. Climatechange and shifts in spring phenology of three horticultural woody perennials in northeasternUSA. Internat. J. Biomet. 49:303–309.

Wu, G.W., and L.T. Wilson. 1997. Growth and yield response of rice to rice water weevil injury.Environ. Entomol. 26:1191–1201.

Zalucki, M.P., and R.D. van Klinken. 2006. Predicting population dynamics of weed biological controlagents: science or gazing into crystal balls? Aust. J. Entomol. 45:331–344.

climate change effects on plant-pest-natural enemy ...nature.berkeley.edu › garbelottowp ›...

Documents