ORIGINAL PAPER

Climate model for seasonal variation in Bemisia tabaci using CLIMEXin tomato crops

Rodrigo Soares Ramos1,2 & Lalit Kumar2 & Farzin Shabani2,3 & Ricardo Siqueira da Silva1 & Tamíris Alves de Araújo1&

Marcelo Coutinho Picanço1

Received: 30 September 2018 /Revised: 24 November 2018 /Accepted: 8 December 2018 /Published online: 24 January 2019# ISB 2019

AbstractThe whitefly, Bemisia tabaci, is considered one of the most important pests for tomato Solanum lycopersicum. The populationdensity of this pest varies throughout the year in response to seasonal variation. Studies of seasonality are important to understandthe ecological dynamics and insect population in crops and help to identify which seasons have the best climatic conditions forthe growth and development of this insect species. In this research, we used CLIMEX to estimate the seasonal abundance of aspecies in relation to climate over time and species geographical distribution. Therefore, this research is designed to infer themechanisms affecting population processes, rather than simply provide an empirical description of field observations based onmatching patterns of meteorological data. In this research, we identified monthly suitability for Bemisia tabaci, with the climatemodels, for 12 commercial tomato crop locations through CLIMEX (version 4.0). We observed that B. tabaci displays season-ality with increased abundance in tomato crops during March, April, May, June, October and November (first year) and duringMarch, April, May, September and October (second year) in all monitored areas. During this period, our model demonstrated astrong agreement between B. tabaci density and CLIMEX weekly growth index (GIw), which indicates significant reliability ofour model results. Our results may be useful to design sampling and control strategies, in periods and locations when there is highsuitability for B. tabaci.

Keywords Seasonality .Whiteflies . Modelling . CLIMEX

Introduction

The whitefly Bemisia tabaci (Gennadius) (Hemiptera:Aleyrodidae) is an extremely polyphagous species, considereda key pest for many crops worldwide, including vegetables,ornamentals and field crops (Gilioli et al. 2014; Naranjo et al.2009; Simmons et al. 2008). The damage caused by this pestmay occur directly through phloem feeding or indirectlythrough the excretion of honeydew and the consequent

development of sooty molds on the leaves, leading to a reduc-tion of photosynthesis, which may consequently lead to plantmortality and production losses of up to 100% (Friedmannet al. 1998; Lapidot et al. 1997; Navas-Castillo et al. 2011;Papayiannis et al. 2010). In Latin America, the tomato(Solanum lycopersicum L.) is one of the main crops attackedby B. tabaci (Morales and Jones 2004; Togni et al. 2010).Additionally, this insect pest transmits several plant viruses.In terms of area coverage, tomatoes suffer one of the greatestcrop losses valued at US$3806 ha−1 caused by insect pests(Oliveira et al. 2014).

Every year, many of papers are published investigatingB. tabaci from different perspectives, such as chemical control(McKenzie et al. 2014), molecular design (Elfekih et al.2018), plant-insect interaction (Han et al. 2016; Ramos et al.2018), disease-insect interaction (Ning et al. 2015), virustransmission (Gottlieb et al. 2010), transcriptome (Xie et al.2012), symbiosis (Luan et al. 2015), resistance to insecticides(Horowitz and Ishaaya 2014), taxonomy (Grávalos et al.2015) and others. However, ecological studies using

* Rodrigo Soares Ramosrodrigosragro@gmail.com; rodrigo.ramos@ufv.br

1 Departamento de Entomologia, Universidade Federal de Viçosa(UFV), Viçosa, MG 36570-900, Brazil

2 Ecosystem Management, School of Environmental and RuralScience, University of New England (UNE), Armidale, NSW 2351,Australia

3 Biological Sciences, Flinders University, GPO Box 2100,Adelaide, South Australia 5001, Australia

International Journal of Biometeorology (2019) 63:281–291https://doi.org/10.1007/s00484-018-01661-2

B. tabaci are frequently neglected due to requiring detailedfield studies over a long period.

B. tabaci is highly sensitive to climate fluctuations due toits ectothermic physiology (Zidon et al. 2016) and exhibitsthese fluctuations as a response to climate variations(Munyuli et al. 2017). Seasonality is a key component to un-derstand the ecology of insect populations in field crops, es-pecially pests (Campos et al. 2006; Pedigo and Rice 2014;Pereira et al. 2007). Ecological studies using insect pest mon-itoring over time in field crops and climate dynamics for insectspecies occurrence provide a useful method to better under-stand the spatio-temporal climate dynamics that determineseasonality patterns of insects in field crops.

One of the tools to determine spatio-temporal climatedynamics for species is using CLIMEX software. It hasbeen considered an inferential modelling software pro-gram, which enables the user to estimate the potential geo-graphical distribution with great reliability (Kriticos et al.2015). This program simulates the mechanisms that limitspecies distribution based on ecophysiological parameters.Additionally, it is possible to describe the spatio-temporaldynamics in climate suitability (Kriticos et al. 2015).Modelling using CLIMEX provides map sequencesdisplaying the suitability changes in both space and time.Thus, it is now possible to increase our understanding ofclimatic influence on spatio-temporal species dynamics(e.g. B. tabaci) and produce robust models that can corre-late with field monitoring data (De Villiers et al. 2016,2013). All this information is relevant to plan managementstrategies and ecological field studies during periods whensuitability and/or insect pest density is higher.

A better understanding of the temporal dynamics offavourable conditions for B. tabaci considering climate dy-namics is needed. Therefore, our aim in this study was todetermine the B. tabaci seasonal variability in a commercialtomato crop over 2 years and the influence of climate dynam-ics. To determine the monthly climatic variations suitable forB. tabaci, we used 12 different locations with tomato cultiva-tion from 2 years (2015 and 2016) of monitoring, as well asthe influence of monthly climate on the species, usingCLIMEX modelling—CLIMEX (version 4.0).

Materials and methods

Bemisia tabaci distribution

In our research, we collected 94 records of Bemisia tabacifrom the published literature (Queiroz et al. 2016, 2017) andthe Global Biodiversity Information Facility (GBIF) and con-firmed these using the software EPPO Plant Quarantine DataRetrieval system (PQR, version 5.3.5, 2015). All records (94

occurrences) used in this study from Central and SouthAmerica are shown in Fig. 1.

Field data collection

In this study, B. tabaci densities were assessed in tomato cropsfrom 12 commercial farms located in Coimbra, Minas GeraisState, Brazil (20° 51′ 24″ S, 42° 48′ 10″ W; altitude 720 m).The tomato crops were established in areas ranging from 2.5to 6 ha, with a spacing of 1.0 m between rows and 0.5 mbetween plants using local procedures (Heuvelink 2005;Jones Jr 2007).

B. tabaci population density was monitored weekly fromthe beginning of the transplanting of the seedlings until the lastharvest for each evaluated tomato crop (a total of 12 crops)over a 2-year period: crop 1, from January 2015 to April 2015;crop 2, from February 2015 to May 2015; crop 3, fromMarch 2015 to June 2015; crop 4, from July 2015 toOctober 2015; crop 5, from November 2015 to December2015; crop 6, from October 2015 to January 2016; crop 7,from December 2015 to February 2016; crop 8, fromFebruary 2016 to March 2016; crop 9, March 2016 toJune 2016; crop 10, from July 2016 to October 2016; crop11, from August 2016 to November 2016; and crop 12, fromOctober 2016 to December 2016. In each field, we evaluatedB. tabaci density in 300 sampling units per location per week.The sampling was performed in all locations, randomly mon-itored on a grid pattern to avoid bias in the choice of samplinglocation (Bacci et al. 2006; da Silva et al. 2017). The samplingmethod applied for B. tabaci involved the direct counting ofadults using plastic trays, where a leaf from the apex of a plantwas struck on a white background tray, and the direct countingof nymphs on a leaf from the basal third of the tomato plant(Gusmao 2000; Gusmão et al. 2005; Imenes et al. 1992; Limaet al. 2017). We counted the number of nymphs present in thebasal part of the tomato canopy and the number of adultspresent on the white plastic tray. Mean temperature, rainfalland photoperiod were obtained daily from a local weatherstation.

CLIMEX model

To study the spatio-temporal dynamics of climate suitability,we chose to use CLIMEX because this software provides aresource to compare locations/years and create maps of aver-age conditions. This tool creates map sequences that allow usto visualize how suitability changes across both space andtime (Kriticos et al. 2015). The simulation of how climaticfactors may simultaneously influence species range in spaceand time is considered a significant and powerful tool notavailable in other software programs (Kriticos et al. 2015).

CLIMEX software simulates the mechanisms that limitspecies’ geographical distributions and determines their

282 Int J Biometeorol (2019) 63:281–291

seasonal phenology, which may affect species abundancesuch as insects (Kriticos et al. 2015). This software enablesus to describe how species respond to climatic variables atdifferent specific times (e.g. daily or weekly) (Kriticoset al. 2015). To set the biological parameters in CLIMEX,we used biological information from the species and fromthe locations where their occurrence is reported.Additionally, this software helps to predict and map poten-tial distributions (Kriticos et al. 2015). The growth andstress indices are combined into an Ecoclimatic Index(EI), which is considered an average yearly index; thisindex gives an overall measure of the climatic suitabilityof a location for a target species. The EI values are scaledfrom 0 to 100.

EI values close to 0 means that the location is inade-quate for the establishment of the species while an EIvalue higher than 30 indicates that there is suitability interms of climatic conditions for the growth and

development of a species (Kriticos et al. 2015). We usedCLIMEX (version 4.0), which is a software program thatcan draw maps that describe the spatial distribution ofspecies over a given period of time using the weeklygrowth index (GIw) and describe suitable conditions forpopulation growth on a scale from 0 to 1 (Kriticos et al.2015). GIw is maximized (GIw > 0) in appropriate sea-sons, when the weekly suitability of climate for speciesgrowth and development is present, and minimized duringunfavourable seasons. To determine the GIw value, weused the temperature (TI) and moisture (MI) indices toinclude B. tabaci growth requirements. We used the stressindices in relation to the moisture index and temperaturefor species survival. These indices need to be consideredbecause all of these abiotic indices contribute to a goodspecies distribution result in relation to the adverse cli-matic conditions that occur at different times throughoutthe year (Kriticos et al. 2015).

Fig. 1 Records of B. tabaci in Central and South America and currentdistribution of B. tabaci in the validation region based on EI. The areas inwhite (EI = 0), yellow (0 < EI < 30) and red (30 < EI < 100) indicate

unfavourable, less favourable and highly favourable climate areas forB. tabaci, respectively. The circle indicates the monitored areas (12different tomato field crops)

Int J Biometeorol (2019) 63:281–291 283

Model calibration and validation

The CLIMEX model for B. tabaci was adjusted using 60occurrence records and biological data of the species, suchas thermal requirements, moisture index and stress indices(cold, heat dry and wet stress). For model validation, datafrom Brazil was omitted, adjusting parameters only. For pa-rameter adjustments, we used B. tabaci biological data fromthe literature and unpublished information from the IntegratedPest Management Lab at Universidade Federal de Viçosa,Minas Gerais, Brazil, where numerous studies with B. tabacihave been conducted in the laboratory and in the field. In ourmodel, we used CliMond 10′–gridded climate data.

To represent historical climate (data from 1961 to 1990,30 years, centred on 1975), we used the average maximummonthly temperature (Tmax), average minimum monthly tem-perature (Tmin), average monthly precipitation (Ptotal) and rel-ative humidity at 0900 hours (RH09:00) and 1500 hours(RH15:00). All values were fitted to match to the locationrecords for the pest with exactness of prediction by theCLIMEX-generated model in different regions (Fig. 1). Theagreement between species distribution and the model for theseasonal phenology provided cross-validation for our model(Kriticos et al. 2015).

Temperature index

Albergaria and Cividanes (2002) have previously reported thethermal requirements for B. tabaci. They presented a lowertemperature threshold of 8.3 °C for life cycle (egg–adult) andan upper temperature threshold of 37 °C. Above this temper-ature, B. tabaci eggs became infertile. We established a lowertemperature limit (DV0) of 8.3 °C and an upper temperaturelimit (DV3) of 37 °C. We considered the range from 15 to35 °C as the most appropriate for B. tabaci survival, growthand development (Albergaria and Cividanes 2002). Regardingthe lower (DV1) and upper (DV2) optimal temperatures, weestablished values of 15 °C and 35 °C, respectively. In thissame research, Albergaria and Cividanes (2002) indicated fora life cycle (egg–adult) value of 472.6 degree days. Thus,PDD was set to 472.6 °C days.

Moisture index

We used B. tabaci records for distribution in wet tropical re-gions with the highest GI values. High B. tabaci incidence intomato cultivation located in Coimbra, Minas Gerais State,Brazil, regularly occurs in seasons with a mean rainfall of90 mm (Leite et al. 2006). For the moisture index, we used0.1 (denoting a permanent wilting point) as a lower soil mois-ture threshold (SM0), 0.4 for lower optimum soil moisture(SM1), 0.7 for upper optimum soil moisture (SM2) and 1.5for an upper soil moisture threshold (SM3). These selections

were used based on the literature and to obtain better modelresults. All adjustments were made in terms of the B. tabacilocation records and considering B. tabaci density within themonitored areas. In addition, these values provided the best fitfor host distribution (for tomato crops).

Stress indices

Cold stress

The cold stress temperature threshold defines a temperaturebelow which cold stress begins to occur. As a poikilothermalspecies, B. tabacimay die when daily thermal accumulation istoo low to maintain metabolism (Kriticos et al. 2015). In an-other words, this occurs when a threshold number of degreedays above the developmental temperature threshold are notreached.

In this context, this parameter is considered an importantparameter for B. tabaci because it affects species survival.Therefore, we determined a threshold number of degree daysin the model, which is defined by the developmental temper-ature threshold (DVCS). This parameter is expressed in coldstress degree day threshold (DTCS) and is defined by degreeday units. In the present study, the DTCS used was 10 °C daysand, for the cold stress degree day, expressed in units per week(DHCS), it was 0 week−1. We selected all these values basedon the published literature which presented some similaritiesfor B. tabaci distribution in Central and South America(Desneux et al. 2010; Gusmão et al. 2006; Leite et al. 2006;Tomar and Malik 2017). All values used to adjust the modelare presented in Table 1 and are matched for the locationswhere B. tabaci occurs.

Heat stress

High temperatures may be excessive and have a negative im-pact on insect growth and development (da Silva et al. 2017;Gerling 1986; Rosenzweig et al. 2001). B. tabaci do not sur-vive temperatures higher than 37 °C (Albergaria andCividanes 2002). The heat stress parameter (TTHS) was setat 35 °C, and its accumulation rate (THHS) was0.0007 week−1. High temperatures may cause physiologicaldisorders in insects and are in accordance with the nonoccur-rence of the species in some regions in northern Brazil.

Dry stress

The most significant, known B. tabaci distributions are re-corded in tropical and subtropical regions with some humidareas (Hirano et al. 1993; Jafarbeigi 2014; Stansly andMcKenzie 2008). Another important parameter that maycause stress for a species is when the soil moisture level istoo low or too high (Kriticos et al. 2015). For our model, the

284 Int J Biometeorol (2019) 63:281–291

soil moisture threshold (SMDS) was set at 0.1 and stress ac-cumulation rate (HDS) at − 0.001 week−1.

Wet stress

Wet stress can affect insects in several ways, especially in-creasing mortality due to high precipitation (Bacci et al.2006; Pereira et al. 2007; Varella et al. 2015). The soil mois-ture threshold for wet stress (SMWS) was set at 2.0, and thestress accumulation rate (HWS) was set at 0.002 week−1. Theparameter values presented an adequate match with knownpest distributions, for example the Brazilian Atlantic Forestarea (da Silva et al. 2017; Gontijo et al. 2013). These valuesare presented in Table 1.

Meteorological data

The monthly time series was used to compare locations byloading the model with the “CL–Grid Data” simulation filewithin a year in CLIMEX (Kriticos et al. 2015). ClimaticResearch Unit (CRU) (CRU TS3.23, Norwich (http://www.cru.uea.ac.uk/cru/data/hrg.htm) Time-Series (TS), version 3.23, with monthly climate variations was used. This versionhas the data for all parameters reformatted which is required

when using this software. Variables such as monthly averagedaily maximum and minimum temperature, precipitation andvapor pressure (Harris and Jones 2017) were included. Then,the CLIMEX model (from January 2015 to December 2016)and maps for the same period during which B. tabaci wasmonitored were generated.

Model validation

The B. tabaci nymph and adult densities observed over 2 yearsof collection in tomato plantations were compared with themonthly suitability maps generated for the same period. Themodel results were found to be consistent with real B. tabacidistribution and therefore demonstrated significant reliability.We used all occurrences registered in Central and SouthAmerica (94 occurrences) to test the model. The percentageof the occurrence data which falls within the model projectionwas calculated and used to evaluate our model’s reliability.

Results

The model presented in this study showed a consistent matchbetween the current distribution of Bemisia tabaci in Central

Table 1 CLIMEX parameter values used for the Bemisia tabaci modelling

Index Parameter Values Unit References

Temperature DV0 = lower threshold 8.3 °C Albergaria and Cividanes (2002)

DV1 = lower optimum temperature 15 °C Albergaria and Cividanes (2002)

DV2 = upper optimum temperature 35 °C Albergaria and Cividanes (2002)

DV3 = upper threshold 37 °C Albergaria and Cividanes (2002)

Moisture SM0 = lower soil moisture threshold 0.1 a Leite et al. (2006)

SM1 = lower optimum soil moisture 0.4 a Leite et al. (2006)

SM2 = upper optimum soil moisture 0.7 a Leite et al. (2006)

SM3 = upper soil moisture threshold 1.5 a Leite et al. (2006)

Cold stress TTCS = temperature threshold 8.3 °C Albergaria and Cividanes (2002)

THCS = stress accumulation rate − 0.001 Week−1 Tomar and Malik (2017); Gusmão et al. (2006)

DTCS = degree day threshold 10 °C days Tomar and Malik (2017); Gusmão et al. (2006)

DHCS = stress accumulation rate – Week−1 Tomar and Malik (2017); Gusmão et al. (2006)

Heat stress TTHS = temperature threshold 35 °C Albergaria and Cividanes (2002)

THHS = stress accumulation rate 0.0007 Week−1 Moraes and Foerster (2015)

Dry stress SMDS = soil moisture threshold 0.1 Jafarbeigi (2014)

HDS = stress accumulation rate − 0.001 Week−1

Wet stress SMWS = soil moisture threshold 2 a

HWS = stress accumulation rate 0.002 Week−1

Hot-wet stress TTHW = hot-wet temperature threshold 32 °C Albergaria and Cividanes (2002)

MTHW = hot-wet moisture threshold 0.2 SMC

PHW = hot-wet stress rate 0.03 Week−1

Degree days PDD = degree days per generation 472.6 °C days Albergaria and Cividanes (2002)

a Values without units are dimensionless indices of soil moisture (0 = over dry, 1 = field capacity)

Int J Biometeorol (2019) 63:281–291 285

and South America and the EI from the CLIMEX (Fig. 1). Tovalidate our model, we checked the reports of occurrence forthis species in all areas (Central and South America), and 99%of Bemisia tabaci occurrences were within favourable climatecategories. This characteristic is important, and we may con-firm the reliability and confidence in the values selected forthe parameters we used to generate our current model by ap-plying the CliMond 10′–gridded climate data for modeling inCLIMEX.

The highest B. tabaci densities were found in crops 2, 3, 6and 10. Of these, crops 3, 6 and 10 showed the highest densityof adults and crop 2 the highest nymph density. The lowestdensity was observed in crops 1, 4, 8, 9, 11 and 12. Only crops5 and 7 showed intermediate density in relation to the othercrops evaluated in this study (Fig. 2).

The highest nymph densities were observed betweenOctober and December in crop 6 in the first year and betweenSeptember and October in crop 10 in the second year. Thepeaks for B. tabaci adults occurred in June: betweenNovember and December for the first year (for crops 3 and6) and in October for the second year (for crop 10) (Fig. 2).

B. tabaci occurrence in crops 1, 4, 8, 9, 11 and 12 (duringJanuary–March and in July until the beginning of September)

was low for B. tabaci nymphs and adults in both years eval-uated. We noticed the highest attack intensity for B. tabaciduring October to December and the lowest intensity duringJanuary to April and during June to September in open-fieldconditions. We observed a spatio-temporal variation for cli-mate suitability for Bemisia tabaci over the 2 years (Fig. 3).October and November were the months with the highestclimate suitability in Central and South America (e.g. coun-tries such as Brazil, Bolivia, Ecuador, Peru, Colombia andAndes Region) (Figs. 4 and 5).

For the greater part of the year, we observed high suitabilityin coastal and southern Brazil and also in countries such asCosta Rica, Guatemala, Honduras, Colombia, Southern Chileand Venezuela. Climate suitability increased in June andOctober–December throughout the five Brazilian regions(north, mid-west, southeast, northeast and south) and largeregions of Bolivia, Peru and Argentina. June, July andAugust present a change to climatically unfavourable condi-tions for B. tabaci in the mid-west and some regions of south-eastern Brazil also. Additionally, a decrease was observed in

Fig. 2 a–d B. tabaci nymph and adult densities in commercial tomatocrops in Coimbra, Minas Gerais, Brazil, in 2015 and 2016

Fig. 3 Variation of temperature and rainfall in the tomato cropsthroughout the experimental period

286 Int J Biometeorol (2019) 63:281–291

January–April in the north and mid-west and in some parts ofsoutheastern Brazil, as well as in other countries such as Peruand Ecuador (Figs. 4 and 5).

When we performed a zoom into an area, we noticed aclimatic variability for B. tabaci that includes the monitoredfields (where the evaluated crops were planted) (Fig. 6).

Increased climatic suitability for B. tabaci was observed be-tween the months of April–May and October–November inthe first year and between the months of March–April andSeptember–October in the second year. The highest growthindex was observed during the months of April–May andOctober–November for the first year and during April and

Fig. 4 Climate variability by month, based on the growth index (0 to 1) for B. tabaci for the Central and South America for year 1. Areas in white colormeans that the growth index is equal to 0 (zero)

Int J Biometeorol (2019) 63:281–291 287

October for the second year. On the other hand, a reductionwas observed between June and September and betweenNovember and January. There was almost zero climate suit-ability for B. tabaci in January and August in both years. InFebruary, climate suitability returns with progressive increasesuntil May–June and September until November (Fig. 6).

Discussion

We noticed that a significant intensity of B. tabaci attacksoccurred during October to December and the lowest intensityoccurred during January to April and June to September inopen-field tomato crops. The main reason for this may be

Fig. 5 Climate variability by month based on the growth index (0 to 1) for B. tabaci for the Central and South America for year 2. Areas in white colormeans that the growth index is equal to 0 (zero)

288 Int J Biometeorol (2019) 63:281–291

related to the environmental conditions of this season. Octoberto December proved to be significant for the growth and de-velopment of the species in open-field conditions because theprecipitation rate is not too high and the temperature duringthis period is within the highest range for the growth anddevelopment of this species. However, this pest species pre-sents great adaptability to different climatic conditions(Sutherst et al. 2011). Additionally, it is common to find ahigh number and diversity of host plants in the field duringthis season, which contributes to the successes of any invasivespecies.

The environment directly and indirectly affects the season-ality of B. tabaci in the field (Alicai 1999; Leite et al. 2006;Naranjo and Ellsworth 2005). The variability of insect densitymay be influenced directly by seasonality due to the impact ofclimate factors (i.e. rain, wind and temperature) and indirectlyby that due to food availability. For B. tabaci, food availabilitydoes not appear to be a main factor when defining B. tabaciseasonality in the field because it is a polyphagous species thatcan be hosted by numerous plant species.

When we see our model results, it is clear that climatesuitability for B. tabaci presented variability in several areasthroughout the months of the year (Figs. 4 and 5). The directinfluence of climatic factors on seasonality for B. tabaci seemsto be greater than the indirect influence of host plant availabil-ity, since host plants (such as tomato crops) were available inthe field for B. tabaci throughout the whole evaluation period.This can be observed in Figs. 2 and 5, where B. tabaci densityvaried over time, coinciding with the highest and lowestgrowth indices, respectively.

B. tabaci density was highest between September andDecember, while its lowest densities were observed duringJanuary–February and June–August. When we observe thedata for B. tabaci density in the field and the CLIMEX modelGIw for these periods, we notice a significant agreement be-tween them, which reinforces confidence in the model’s re-sults. Further, the rapid decrease and increase observed duringthe evaluation period in the field matched the period of time(an increase and decrease of B. tabaci densities with thegrowth index of the CLIMEX model) which reinforces thevalidity, highlighting the robustness of the model presented.

In the second year of evaluation, B. tabaci densities intomato crops were low from February to May but our modelshowed a progressive increase in the growth index for the areamonitored. One of the possible reasons that might explain thisfact is related to the effectiveness of the pest control methodapplied by the farmers in these crops, which contributes to thelow B. tabaci density. Additionally, this was a period whererainfall was very low, which affected the planting of manyvegetable crops (including tomato cultivation) in the region.Most of the host crops require a significant volume of waterfor growth and development; in cases of adverse conditions,this implies a low number and diversity of hosts (low foodavailability). Consequently, the population of this pest in thefield over the colonization period was very low. This is justone example of how climate conditions are important and mayindirectly affect species fluctuations. In the present case, thismay have affected the environment and, consequently,B. tabaci density. In addition, the climate may affect naturalenemy populations, leading to low density which implies a

Fig. 6 Climate variability by month based on the growth index (0 to 1) for B. tabaci, displaying an area and the location of the 12 monitored tomatocrops. Areas in white color means that the growth index is equal to 0 (zero)

Int J Biometeorol (2019) 63:281–291 289

low rate of natural biological control. Therefore, it is importantto intensify the monitoring at this time of the year since theclimate conditions show themselves to be favourable.

The combination of a high number of host plants (tomatoplantations), favourable climate, the presence of B. tabaci andinadequate control methods in the field led to increasedB. tabaci numbers in the monitored area over 5–6 months ofthe year.

In March, from the first year, we observed initial coloniza-tion by nymphs in the monitored areas (Fig. 2). Even thoughthis period was favourable for B. tabaci, no increase in thenumber of adults was observed. This might be explained bythe efficient control applied during that stage, which contrib-uted to decreased nymph numbers, thereby reducing the num-ber of individuals that reached adult stage.

In October (in the first year) and September (in the secondyear), colonization by B. tabaci started in the monitored areasand a population increase (nymphs and adults) on crops wasobserved over time. This occurred because no efficient controlmethod was applied and all conditions (favourable climate,food availability and pest presence) were significant duringthis period.

This seasonal variation matches our spatio-temporal cli-mate dynamics model. These results provide a significant con-tribution for management and a further study of B. tabaci infield crops, since the model results determine seasons withfavourable climate conditions for occurrence and the best mo-ment to encounter B. tabaci. Our results indicate periods whena risk of B. tabaci is highest and may help farmers to focus onspecific times of the year for B. tabaci control. Additionally,further research could be conducted to determine other impor-tant factors (e.g. natural enemies) that influence the seasonalvariations of B. tabaci and its interactions with spatio-temporal climate dynamics in terms of ecological studies.

Acknowledgements The simulations were carried out using the compu-tational facilities at UNE. Mr. Phillip John Villani (B.A. from theUniversity of Melbourne, Australia) revised and corrected the Englishlanguage used in this manuscript.

Author contributions RSR, RSS and MCP conceived of and designedthe research. TAA, RSS and RSR contributed to conducting the experi-ments and acquiring the data. RSR analysed the data and wrote the man-uscript with support from LK. LK and FS made the critical revisions(providing language help and writing assistance). LK and MCP madethe critical revisions and approved the final version. All authors reviewedand approved the final manuscript.

Funding information This research was supported by the NationalCouncil for Scientific and Technological Development (ConselhoNacional de Desenvolvimento Científico e Tecnológico (CNPq)) andfinanced in part by the Coordenação de Aperfeiçoamento de Pessoal deNível Superior (CAPES) of Brazil (Finance Code 001), the Minas GeraisState Foundation for Research Aid (FAPEMIG) and the School ofEnvironmental and Rural Science of the University of New England(UNE), Armidale, Australia.

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict ofinterest.

References

Albergaria NM, Cividanes FJ (2002) Thermal requirements of Bemisiatabaci (Genn.) B-biotype (Hemiptera: Aleyrodidae). NeotropEntomol 31(3):359–363

Alicai T (1999) Seasonal changes in whitefly numbers and their influence onincidence of sweetpotato chlorotic stunt virus and sweetpotato virusdisease in sweetpotato in Uganda. Int. J. Pest Manag. 45(1):51–55

Bacci L, Picanço MC, Moura MF, Della Lucia TM, Semeão AA (2006)Sampling plan for Diaphania spp.(Lepidoptera: Pyralidae) and for hy-menopteran parasitoids on cucumber. J EconEntomol 99(6):2177–2184

Campos WG, Schoereder JH, DeSouza OF (2006) Seasonality in neo-tropical populations of Plutella xylostella (Lepidoptera): resourceavailability and migration. Popul Ecol 48(2):151–158

da Silva RS,Kumar L, Shabani F, da Silva EM, da SilvaGaldino TV, PicançoMC (2017) Spatio-temporal dynamic climate model for Neoleucinodeselegantalis using CLIMEX. Int J Biometeorol 61(5):785–795

De Villiers M, Hattingh V, Kriticos D (2013) Combining field phenolog-ical observations with distribution data to model the potential distri-bution of the fruit fly Ceratitis rosa Karsch (Diptera: Tephritidae).Bull Entomol Res 103(1):60–73

De Villiers M, Hattingh V, Kriticos DJ, Brunel S, Vayssières J-F,Sinzogan A, Billah M, Mohamed S, Mwatawala M, Abdelgader H(2016) The potential distribution of Bactrocera dorsalis: consideringphenology and irrigation patterns. Bull Entomol Res 106(1):19–33

Desneux N, Wajnberg E, Wyckhuys KA, Burgio G, Arpaia S, Narváez-Vasquez CA, González-Cabrera J, Ruescas DC, Tabone E, FrandonJ (2010) Biological invasion of European tomato crops by Tutaabsoluta: ecology, geographic expansion and prospects for biologi-cal control. J Pest Sci 83(3):197–215

Elfekih S, Tay WT, Gordon K, Court LN, De Barro PJ (2018)Standardizedmolecular diagnostic tool for the identification of cryp-tic species within the Bemisia tabaci complex. Pest Manag Sci74(1):170–173

FriedmannM, Lapidot M, Cohen S, PilowskyM (1998) A novel source ofresistance to tomato yellow leaf curl virus exhibiting a symptomlessreaction to viral infection. J Am Soc Hortic Sci 123(6):1004–1007

Gerling D (1986) Natural enemies of Bemisia tabaci, biological charac-teristics and potential as biological control agents: a review. AgricEcosyst Environ 17(1):99–110. https://doi.org/10.1016/0167-8809(86)90031-9

Gilioli G, Pasquali S, Parisi S, Winter S (2014) Modelling the potentialdistribution of Bemisia tabaci in Europe in light of the climatechange scenario. Pest Manag Sci 70(10):1611–1623

Gontijo P, Picanço M, Pereira E, Martins J, Chediak M, Guedes R (2013)Spatial and temporal variation in the control failure likelihood of thetomato leaf miner, Tuta absoluta. Ann Appl Biol 162(1):50–59

Gottlieb Y, Zchori-Fein E, Mozes-Daube N, Kontsedalov S, Skaljac M,Brumin M, Sobol I, Czosnek H, Vavre F, Fleury F (2010) Thetransmission efficiency of tomato yellow leaf curl virus by thewhite-fly Bemisia tabaci is correlated with the presence of a specific sym-biotic bacterium species. J Virol 84(18):9310–9317

Grávalos C, Fernández E, Belando A, Moreno I, Ros C, Bielza P (2015)Cross-resistance and baseline susceptibility of Mediterranean strains ofBemisia tabaci to cyantraniliprole. Pest Manag Sci 71(7):1030–1036

Gusmão M, Picanço M, Guedes R, Galvan T, Pereira E (2006) Economicinjury level and sequential sampling plan for Bemisia tabaci in out-door tomato. J Appl Entomol 130(3):160–166

290 Int J Biometeorol (2019) 63:281–291

Gusmao MR (2000) Avaliação de vetores de viroses, predadores eparasitóides e plano de amostragem para mosca-branca emtomateiro. Universidade Federal de Viçosa

Gusmão MR, Picanço MC, Zanuncio JC, Silva DJH, Barrigossi JAF(2005) Standardised sampling plan for Bemisia tabaci(Homoptera: Aleyrodidae) in outdoor tomatoes. Sci Hortic 103(4):403–412. https://doi.org/10.1016/j.scienta.2004.04.005

Han P, Desneux N, Michel T, Le Bot J, Seassau A, Wajnberg E, Amiens-Desneux E, Lavoir A-V (2016) Does plant cultivar difference mod-ify the bottom-up effects of resource limitation on plant-insect her-bivore interactions? J Chem Ecol 42(12):1293–1303

Harris I, Jones P (2017) CRU TS4. 00: Climatic Research Unit (CRU)Time-Series (TS) version 4.00 of high resolution gridded data ofmonth-by-month variation in climate (Jan. 1901–Dec. 2015).Centre for Environmental Data Analysis 25

Heuvelink E (2005) Tomatoes, vol 13. CABI,Hirano K, Budiyanto E, Winarni S (1993) Biological characteristics and

forecasting outbreaks of the whitefly, Bemisia tabaci, a vector ofvirus diseases in soybean fields. ASPAC Food & FertilizerTechnology Center

Horowitz AR, Ishaaya I (2014) Dynamics of biotypes B and Q of thewhitefly Bemisia tabaci and its impact on insecticide resistance. PestManag Sci 70(10):1568–1572

Imenes S, Campos T, Takematsu A, Bergmann E, Silva M (1992) Efeitodo manejo integrado na população de pragas e inimigos naturais naprodução de tomate estaqueado. Arq Inst Biol 59:1–7

Jafarbeigi F (2014) Sublethal effects of some botanical and chemicalinsecticides on the cotton whitefly, Bemisia tabaci (Hem:Aleyrodidae). Arthropods 3(3):127

Jones JB Jr (2007) Tomato plant culture: in the field, greenhouse, andhome garden. CRC

Kriticos DJ, Maywald GF, Yonow T, Zurcher EJ, Herrmann NI, SutherstR (2015) Exploring the effects of climate on plants, animals anddiseases. CLIMEX Version 4:184

Lapidot M, Friedmann M, Lachman O, Yehezkel A, Nahon S, Cohen S,PilowskyM (1997) Comparison of resistance level to tomato yellowleaf curl virus among commercial cultivars and breeding lines. PlantDis 81(12):1425–1428

Leite GLD, Picanço M, Guedes RNC, Ecole CC (2006) Factors affectingthe attack rate of Bemisia tabaci on cucumber. Pesq Agrop Brasileira41(8):1241–1245

LimaCH, Sarmento RA, Pereira PS, Galdino TV, Santos FA, Silva J, PicançoMC (2017) Feasible sampling plan for Bemisia tabaci control decision-making in watermelon fields. Pest Manag Sci 73:2345–2352

Luan J-B, Chen W, Hasegawa DK, Simmons AM, Wintermantel WM,Ling K-S, Fei Z, Liu S-S, Douglas AE (2015) Metabolic coevolu-tion in the bacterial symbiosis of whiteflies and related plant sap-feeding insects. Genome Biol Evol 7(9):2635–2647

McKenzie CL, Kumar V, Palmer CL, Oetting RD, Osborne LS (2014)Chemical class rotations for control of Bemisia tabaci (Hemiptera:Aleyrodidae) on poinsettia and their effect on cryptic species popu-lation composition. Pest Manag Sci 70(10):1573–1587

Moraes CP, Foerster LA (2015) Thermal requirements, fertility, and numberof generations of Neoleucinodes elegantalis (Guenée) (Lepidoptera:Crambidae). Neotrop Entomol:1–7 GBIF.org (2nd May 2017) GBIFOccurrence Download https://doi.org/10.15468/dl.mwb31k

Morales FJ, Jones PG (2004) The ecology and epidemiology of whitefly-transmitted viruses in Latin America. Virus Res 100(1):57–65https://doi.org/10.1016/j.virusres.2003.12.014

Munyuli T, Kalimba Y, Mulangane EK, Mukadi TT, Ilunga MT, MukendiRT (2017) Interaction of the fluctuation of the population density ofsweet potato pests with changes in farming practices, climate and phys-ical environments: a 11-year preliminary observation from South-KivuProvince, Eastern DRCongo. Open Agriculture 2(1):495–530

Naranjo SE, Castle SJ, De Barro PJ, Liu S-S (2009) Population dynamics,demography, dispersal and spread of Bemisia tabaci. In: Bemisia:bionomics and management of a global pest. Springer, pp 185–226

Naranjo SE, Ellsworth PC (2005) Mortality dynamics and populationregulation in Bemisia tabaci. Entomol Exp Appl 116(2):93–108

Navas-Castillo J, Fiallo-Olivé E, Sánchez-Campos S (2011) Emergingvirus diseases transmitted by whiteflies. Annu Rev Phytopathol49:219–248

NingW, Shi X, Liu B, PanH,WeiW, ZengY, SunX, XieW,Wang S,WuQ (2015) Transmission of tomato yellow leaf curl virus by Bemisiatabaci as affected by whitefly sex and biotype. Sci Rep 5:10744

Oliveira C, Auad A, Mendes S, Frizzas M (2014) Crop losses and theeconomic impact of insect pests on Brazilian agriculture. Crop Prot56:50–54

Papayiannis LC, Iacovides TA, Katis N, Brown J (2010) Differentiation oftomato yellow leaf curl virus and tomato yellow leaf curl Sardinia virususing real-time TaqMan® PCR. J Virol Methods 165(2):238–245

Pedigo LP, Rice ME (2014) Entomology and pest management.Waveland

Pereira E, Picanço M, Bacci L, Crespo A, Guedes R (2007) Seasonalmortality factors of the coffee leafminer, Leucoptera coffeella. BullEntomol Res 97(4):421–432

Queiroz PR, Lima LH, Sujii ER, Monnerat RG (2017) Description of themolecular profiles of Bemisia tabaci (Hemiptera: Aleyrodidae) indifferent crops and locations in Brazil. Journal of Entomology andNematology 9(5):36–45

Queiroz PR, Martins ES, Klautau N, Lima L, Praça L, Monnerat RG(2016) Identification of the B, Q, and native Brazilian biotypes ofthe Bemisia tabaci species complex using scar markers. Pesq AgropBrasileira 51(5):555–562

Ramos RS, Kumar L, Shabani F, PicançoMC (2018)Mapping global risklevels of Bemisia tabaci in areas of suitability for open field tomatocultivation under current and future climates. PLoS One 13(6):e0198925. https://doi.org/10.1371/journal.pone.0198925

Rosenzweig C, Iglesias A, Yang X, Epstein PR, Chivian E (2001) Climatechange and extreme weather events; implications for food production,plant diseases, and pests. Glob. Chang. Hum. Health 2 (2):90–104

Simmons AM, Harrison HF, LING KS (2008) Forty-nine new host plantspecies for Bemisia tabaci (Hemiptera: Aleyrodidae). Entomol Sci11(4):385–390

Stansly PA,McKenzie CL (2008) Fourth international Bemisia workshopinternational whitefly genomics workshop December 3–8, 2006,Duck Key, Florida, USA. J Insect Sci 8(4):1–54

Sutherst RW, Constable F, Finlay KJ, Harrington R, Luck J, Zalucki MP(2011) Adapting to crop pest and pathogen risks under a changingclimate. Wiley Interdiscip Rev Clim Chang 2(2):220–237

Togni PH, Laumann RA, Medeiros MA, Sujii ER (2010) Odour maskingof tomato volatiles by coriander volatiles in host plant selection ofBemisia tabaci biotype B. Entomol Exp Appl 136(2):164–173

Tomar S, Malik SSK (2017) Life parameters of whitefly (Bemisia tabaci,Genn.) on different host plants. Indian J Sci Res 16(1):34–37

Varella AC, Menezes-Netto AC, de Souza Alonso JD, Caixeta DF,Peterson RK, Fernandes OA (2015) Mortality dynamics ofSpodoptera frugiperda (Lepidoptera: Noctuidae) immatures inmaize. PLoS One 10(6):e0130437

XieW, Q-sM, Q-jW, S-lW, Yang X, N-n Y, Li R-m, X-g J, H-p P, B-m L(2012) Pyrosequencing the Bemisia tabaci transcriptome reveals ahighly diverse bacterial community and a robust system for insecti-cide resistance. PLoS One 7(4):e35181

Zidon R, Tsueda H, Morin E, Morin S (2016) Projecting pest populationdynamics under global warming: the combined effect of inter-andintra-annual variations. Ecol Appl 26(4):1198–1210

Int J Biometeorol (2019) 63:281–291 291