Introduction

Sugarcane is the main source of raw material for Ecuador’s sugar agroindustry and represents the country’s second largest permanent crop in terms of production volume. In 2022, a total of 116,515 ha were planted nationwide, of which 113,148 ha were harvested, yielding 7,740,492 Tm of production. By 2023, these figures declined to 79,580 ha planted and 75,222 ha harvested - a 33.5% decrease – resulting in a 6,253,732 Tm of production. Sugarcane plantations are primarily concentrated in Guayas1. 70% of the cultivated area is located in the lower Guayas Basin, encompassing regions such as Milagro, Naranjito, Marcelino Maridueña (Ingenio San Carlos), El Triunfo and Playas.

In Ecuador, Perkinsiella saccharicida (Homoptera, Delphacidae) is a sucking insect that particularly affects sugarcane crops and is present across all sugarcane-growing regions of the country. It causes substantial yield losses ranging from12 to 36.68 tons per hectare2. However, the direct damage caused by the leafhopper through its bites on the sugarcane foliage is not considered the most significant threat. The primary concern is that this insect is a vector of the Fiji disease virus, a highly dangerous disease that is not currently present in the Americas, is difficult to control, and can cause severe crop losses. Therefore, preventive control of P. saccharicida3 is of critical importance.

According to Pulido 1980, earlier studies in Mauritius in 1903 and 1975 found that the pest’s phenological stage and duration are influenced by temperature and relative humidity. At temperatures of 22ºC and 25ºC, the life cycle of Perkinisella lasted 48 days and 56 days, respectively. The egg incubation stage can last 14 days, but at temperatures below 20ºC it can last 35 to 40 days4.

Scientific studies conducted worldwide indicate that climate change is a reality5,6,7,8. This phenomenon has contributed to rising temperatures, resulting in brief but abundant rainy seasons in some areas and prolonged dry seasons in others9. These changes affect the dissemination and distribution mechanisms of insects, and Perkinsiella is no exception. Temperature is the environmental factor that exerts the most significant influence on insect development10,11,12,13,14, as it affects development rates, survival, fecundity, and dispersion of several species15,16. According to Challinor et al.17, by 2030, crop yields are expected to decline by up to 70%, with losses ranging from 10 to 50% in approximately half of all cases due to climate change. Similarly, Savary et al.18, estimates that between 10% and 28% of global crop production is lost due to pests. This study aimed to evaluate the presence of P. saccharicida in sugarcane crops and analyze its relationship with environmental variables to determine whether climatic variations during both winter and summer influence its population dynamics.

Materials and methods

Study area

Ecuador is located in the southwest of South America. Its climate is influenced by oceanographic factors, atmospheric circulation and marine currents attributable to the Intertropical Convergence Zone (ITCZ)19. The country’s climate variability is attributed to the El Niño-Southern Oscillation (ENSO), with El Niño events characterized by abnormally warm ocean temperatures and increased cloudiness, and precipitation, while La Niña events are associated with abnormally cold ocean temperatures, leading to reduced cloudiness and precipitation. This allows the country to have four different climate regions. The coastal region characterizes for having a tropical climate20,21,22.

This study was conducted in Marcelino Maridueña, located 65 km away from Guayaquil, at a geographical position of 02°15’52’’ South Latitude and 79°28’00’’ West Longitude, and an altitude of 80 m.a.s.l. (Fig. 1). The area experiences a maximum temperature of 30.8 °C and a minimum of 22.3 °C, with a thermal oscillation of 8.5 °C. The potential evaporation is 1,027 mm, with 784 h per year of sunshine and 2,699 MJ*m2*year of solar radiation23.

Fig. 1
figure 1

Source: Own elaboration.

Location of study area. Lots infested with Perkinsiella are highlighted in blue.

Full size image

Sugarcane population and sample size

Sugar cane belongs to the grass family and to the species Saccharum officinarum L. The size of the sugarcane population was of 12,169.53 hectares, distributed across 890 lots ranging from 3 to 40 hectares. Out of these, 96 lots (totaling 1,246 hectares) were evaluated. Observations were conducted annually from 2006 to September 2023, following the recommendations of the Entomology and the Sugar Cane Research Center of Ecuador (CINCAE).

The evaluation involved randomly selecting 50 sugarcane shoots per lot, each with up to 90 days of growth. For each shoot, the number of adult insects, as well as large and small nymphs (nymphs close to adulthood), were counted. It should be noted though, that the values presented in the following section were calculated as the average number of insects per shoot. To extrapolate them to hectares, they were multiplied by 70,000 (\(\:70\:rows*\left(10\:shoots*100\:mt\right),\) which corresponds to the number of shoots per hectare.

The lots with the highest insect infestation were monitored over a ten-year period (2006–2016). The results showed an annual average yield of 79.84 tons of sugarcane per hectare (ha), compared to a production potential of 97.1792 tons per hectare (based on the average of five harvests). This resulted in 17.33 tons of losses (sugarcane that was not harvested) per ha, resulting in an annual average of 17.83% of tons lost.

Description of variables

The study was based on the analysis of Perkinsiella abundance in relation to climatic variables. The data used in this study was obtained from the records made by officials from the Entomology and Statistics (Perkinsiella) and Meteorological (climate variables obtained from the “Casa Blanca” weather station located within San Carlos) Departments of Ingenio San Carlos. The insect population data was compared with the climate variables. Table 1 provides a detailed overview of these variables.

Table 1 Description of variables.
Full size table

Source: Meteorological (Agronomy department) and insect data (Entomology and Statistics department) from Ingenio San Carlos.

Climatology

To demonstrate that considerable climatic variations have occurred in the study area, climatological analyses were performed. The current Technical Regulation No. 49 of the World Meteorological Organization (WMO)24 states that climatological normals correspond to the mean of the climatological data over a 30-year period, which starts in a year ending in one and finishes in a year ending in zero. These normals should be updated every ten years.

To calculate the current climatological normal, historical records of the climatic parameters from the past 30 years (a retrospective series starting from the most recent year ending in zero − 2020) were obtained from the Casa Blanca weather station. This period corresponded to the years 1991 to 2020, in accordance with WMO24 guidelines. As a first step, for each atmospheric parameter, the climatological normal (1991–2020) was calculated as the monthly average across the 30-year period, along with the corresponding monthly standard deviations. Subsequently, for the 2006–2023 data series, anomalies were calculated as the difference between the actual recorded value in a given month and the corresponding monthly mean value for that same month (climatological normal, 1991–2020). This produced a new series of anomalies for the period 2006–2023.

For this study, these anomalies were then normalized by dividing each monthly anomaly by the standard deviation corresponding to that month (calculated similarly to the climatological normal 1991–2020). This method of normalizing anomalies is used to reduce variability due to the long-term climatic fluctuations that has been observed. According to Trewin25, “for the most part, large-scale climatic fluctuations consist of nonlinear variations that oscillate, in the long term, irregularly around a climatological mean value.“. On the other hand, Perkinsiella anomalies (2006–2023), were calculated as the difference between the recorded values and the corresponding monthly means and subsequently normalized by dividing each value by the monthly standard deviation for the same period. These anomalies, were then used for analyses between Perkinsiella and atmospheric parameters.

Weather seasons

The analysis considered two typical weather seasons of the coastal region: the dry-summer season, which includes the months of low precipitation (June to November), and the rainy-winter season (December to May)19,20.

Statistical methods

Interpolation of data

In the case of Perkinsiella, atypical values were found in the months of July and August 2011.

The data series exhibited a sinusoidal pattern (behavior similar to the sine-cosine functions due to their association with the climatic cycles). Hence, the shape preserving interpolant method for curves known as Piecewise Cubic Hermite Interpolating Polynomial (PCHIP) was employed for interpolation to preserve the monotonicity and convexity, present in the data26. This technique uses cubic polynomials defined by segments, with the objective of estimating-interpolating values ​​between known data points. The process begins by dividing the data into intervals and then calculating the derivatives at the data points so that the smoothness of the curve is maintained. Subsequently, cubic polynomials are constructed for each interval, so that the curve passes through the points while preserving consistent derivatives. This approach, which fits a specific cubic polynomial between each pair of points, was found to provide the most accurate interpolation for the weather parameters used in this study27.

Statistical analysis

Descriptive statistical analyses were conducted, including measures of central tendency (mean and median) and dispersion (standard deviation, range, coefficient of variation, and standard error of the mean).

Boxplot graphs were used to identify outliers, and time series graphs were employed to reveal behavioral patterns. A convolution method filter, MA (3), was applied to enhance and highlight the presence of intense events in the time series. Non-parametric hypothesis tests were employed, considering factors such as non-normality, small sample sizes, and the presence of outliers, with a significance level of \(\:\alpha\:=0.05\). Lilliefors28 was used to test for normality in big samples (n > = 50), and Shapiro29 for small samples (n < 50). To test for homogeneity of variance, Bartlett’s test was applied when assuming normality, whereas Levene’s test was employed in cases of non-normality.

For comparing means across more than two independent groups (e.g., year–season combinations), ANOVA was applied along with TukeyHSD when variances were homogeneous.When variances were not homogeneous, equality of medians Kruskall Wallis test30 and Wilcoxon’s pairwise31 were applied. Spearman test32 was used to study the linear relationship between the insect population and climate. Cross-correlation function (CCF) was used to identify out-of-phase correlations. To test for causality in the time series, the Granger test33,34 was applied to determine if one time series can forecast another time series. Stationarity was evaluated using the Dickey–Fuller test35. Cointegration analysis was conducted using the Johansen test, and a Vector Error Correction Model (VECM) was applied, given that the series were cointegrated and exhibited a long-term equilibrium relationship. The statistical analyses and map were performed using RStudio 2024.12.1 and QGIS 3.40.1 software.

Results

Analysis of atmospheric parameters anomalies based in the current climatological normal, the most recent atmospheric behavior in the coastal region (Ecuadorian coastal region), severity of pest infestation, and the relationship between climate and Perkinsiella abundance are presented. The period 2006–2023 was considered for all analyses throughout this research, prioritizing the continuity of complete data in the variables of interest.

Atmospheric conditions of the study area

The atmospheric behavior from January 2006 to September 2023 showed different trends depending on the climatic season (Table 2).

Table 2 Atmospheric conditions and Perkinsiella by weather season.
Full size table

Period: 2006–2023, VC = variation coefficient, s.e.=standard error.

Rainy season

During this season, climatic variables exhibited the highest values for mean, median, minimum, and maximum. Ther_O, Evap and Helio showed greater variability, as indicated by their higher ranges and coefficients of variation. Standard errors (s.e.) were also higher for Acum_P, Evap, Helio, and Solar_R, indicating greater uncertainty in the mean values recorded during this time of the year (Table 2).

Dry season

Ther_O showed a similar mean to that of the rainy season, Evap and Helio reached higher minimum values, Acum_P was more variable (CV = 379.4%), as well as Helio (Table 2).

During the analysis period, the study area experienced notable climatic events, whose effects on atmospheric parameters are evident in the time series of anomalies. Starting in 2010, an upward trend was observed in temperature, Acum_P, and Solar_R (Fig. 2a, b, c, d, h), with the anomaly series (orange line) fluctuating above the central baseline (0). In contrast, Evap, Ther_O, and Helio (Fig. 2e, f, g) showed a downward trend, indicating a general decrease in these parameters over time (Fig. 2).

Fig. 2
figure 2

Standardized anomalies according to the climatological normal of 1991–2020. Trajectory of the series, adjusted to a non-linear Loess model (in orange).

Full size image

Notable positive peaks were observed across all analyzed climatic parameters, most prominently in temperature and Acum_P. These peaks were especially evident during the periods 2014–2016, 2018–2019, and 2023, indicating that during these years, the recorded values in the study area exceeded the expected climatological normals based on the most recent 30-year period. Similarly, negative anomalies with pronounced troughs were detected during 2007–2008, 2010–2012, 2016–2018, and 2020–2023, reflecting periods when climatic values fell significantly below the long-term average.

Severity of Perkinsiella infestation

Initially, Perkinsiella was scarcely present, with several months showing no recorded individuals. Thereafter, it exhibited both high and low abundances. The highest peaks occurred in July 2011 and 2020, with average insect counts per outbreak reaching 58.51 and 56.40, respectively (Fig. 3a). Seasonal changes in Perkinsiella behavior were evident. The insect was more abundant during the rainy season; however, since 2019, a decrease in median counts was observed, and the combined values of nymphs and adults became more consistent, reflecting reduced variability. The dry season was characterized by low counts, but since 2011, an increase in population size was noted along with greater variability and the presence of outliers, registering higher maximum values (Fig. 3b).

Fig. 3
figure 3

Perkinsiella abundance to the period 2006–2023. (a) Time series and (b) Boxplot by weather season.

Full size image

The seasonal change in patterns in both the insect and climate was analyzed in greater detail. Between 2006 and 2010, the insect exhibited distinct seasonal behavior, showing its highest abundance only during the rainy season (Fig. 4a). However, from 2011 to 2023, this pattern shifted, with a higher average abundance from March to October, in both rainy and dry seasons (Fig. 4b).

Fig. 4
figure 4

Perkinsiella monthly abundance (a) 2006–2010 and (b) 2011–2023. Means are denoted as blue points; dashed line illustrates the monthly evolution of the abundance mean.

Full size image

Perkinsiella-climate relationship

The standardized anomalies for the period 2006–2023 were analyzed to study the insect-climate relationship, with an emphasis on differences between weather seasons.

Rainy season

climatic variables exhibited higher mean, median, minimum and maximum values, particularly in temperature, Acum_P, and Solar_R. During this season, Perkinsiella reached a peak average abundance of 28.36 insects per outbreak (Table 2).

Dry Season: Perkinsiella abundance peaked at 58.51 insects per outbreak on average, the highest values observed across all sample statistics. Min_T and Acum_P showed higher coefficients of variation during this season (Table 2).

Across both seasons, fluctuations were observed in most climatic variables, except for Solar_R, where increases occurred during the rainy season and decreases during the dry season. To analyze the relationship between Perkinsiella and atmospheric parameters, Spearman correlation coefficients were calculated due to the presence of outliers (evident as peaks in the anomaly series, Fig. 2) and non-normal distributions—except for Max_T, Ther_O, and Solar_R, which met normality assumptions (Table 3).

Table 3 Normal test to Perkinsiella and atmospheric parameters.
Full size table

Despite the observed changes in both Perkinsiella and climatic conditions, statistically significant linear correlations were identified by weather season (Figs. 5 and 6). Significant correlation coefficients (p-value < 0.05) were obtained for the rainy season between Perkinsiella and Min_T (coef. correlation=−0.36) and Helio (coef. correlation = 0.24). Additionally, Med_T (coef. correlation = −0.19) and Ther_O (coef. correlation = 0.16) were significant at the α = 0.10 level. Positive correlations were observed with Ther_O and Helio, while negative correlations were observed with Min_T and Med_T, suggesting that Perkinsiella is particularly sensitive to increases in minimum temperatures (Fig. 5).

Temperature-related parameters exhibited significant positive linear correlations with the other atmospheric parameters (p-values < 0.05) in both climatic seasons, although stronger correlations occurred during the rainy season (Table 3; Figs. 5 and 6).

Fig. 5
figure 5

Spearman correlation to Rainy season, statically significant values (***, **, *,.) with α (0, 0.01, 0.05, 0.10) respectively.

Full size image
Fig. 6
figure 6

Spearman correlation to Dry season, statically significant values (***, **, *,.) with α (0, 0.01, 0.05, 0.10) respectively.

Full size image

Figures 5 and 6 showed a weak linear trend between the climatic parameters and Perkinsiella, consistent with the low correlation coefficients, some of which were statistically significant.

The significant relationship between Perkinsiella and atmospheric parameters was further analyzed using the CCF, applied to the complete time series to fulfill the continuity condition required for a time series analysis. As shown in Fig. 7, it is evident that the strongest significant correlation (above the confidence limit) between the insect and climatic parameters occurred at different time lags. These correlations were positive in all atmospheric parameters except for Acum_P, indicating that above-average values in atmospheric variables are associated with a subsequent increase in Perkinsiella abundance. high Perkinsiella abundance was observed approximately six months after elevated values of Max_T, Med_T, and Solar_R. In contrast, for Evap, Ther_O, and Helio, high insect abundance preceded the increase in these climatic parameters by one month. In the case of Min_T and Acum_P, the strongest associations with Perkinsiella occurred at lags greater than 12 months.

Fig. 7
figure 7

CCF between atmospheric parameters (x + h) and Perkinsiella (y).

Full size image

To assess whether the insect–climate relationships exhibited causality, the Granger causality test (for stationary time series) and the Toda–Yamamoto approach (for cointegration analysis in non-stationary time series) were applied.

Table 4 Causality Granger test (stationary series).
Full size table

To perform the bivariate Granger test, time series for Perkinsiella and the atmospheric variables Min_T, Acum_P, Evap, Ther_O, and Helio were used. These series satisfied the stationarity condition, as confirmed by the Dickey–Fuller test (p-values < 0.05). None of the atmospheric parameters showed statistically significant Granger causality on Perkinsiella (p-values > 0.05); however, it should be noted that Min_T showed a p-value of 0.12, which is only slightly above the significance threshold of α = 0.10 (Table 4).

To further investigate potential causality, we explored whether joint dynamic interactions among multiple atmospheric variables could explain Perkinsiella dynamics.

Table 5 Cointegration test.
Full size table

Table 5 shows that only one relationship failed to demonstrate cointegration among the variables, as the null hypothesis was not rejected (9.88 < 11.65) at a 1% significance level. However, eight relationships exhibited cointegration, justifying the application of a Vector Error Correction Model (VECM). The model was implemented with a cointegration rank of 6, indicating that up to six cointegration relationships were estimated.

The VECM results revealed that the first lag ((values ​​from the previous month) of Perkinsiella, Max_T, Evap, Helio, and Solar_R were statistically significant predictors of Perkinsiella abundance, with the model explaining 34.3% of the variance. The error correction term (Ect1) was negative, indicating that the system tends to return to equilibrium following a disturbance (Table 6).

Table 6 VECM model (multivariant).
Full size table

Based on the VECM model results (Table 6), it is estimated that a 1 °C increase in Max_T leads to an increase of approximately 1.17 Perkinsiella individuals. Similarly, a one-unit increase in Helio is associated with an increase of 0.34 individuals. In contrast, a one-unit increase in Evap and Solar_R is associated with decreases of 0.34 and 0.27 Perkinsiella individuals, respectively.

Fig. 8
figure 8

Perkinsiella forecast. Historical data 01/2006–09/2023 (cyan) and forecast 10/2023-05/2024 (red).

Full size image

The forecasts revealed the typical upward trend in Perkinsiella abundance during the winter months, with the highest predicted values occurring between April and August (as shown in the abundance forecast column). The VECM model, calculated the forecast in the form of standardized anomalies; these values were converted back to abundance values by applying the inverse of the standardization process (Table 7).

Table 7 Perkinsiella forecast– 12 months period.
Full size table

To explain the observed changes in Perkinsiella abundance and assess the influence of climatic variables, the non-parametric Kruskal–Wallis test was applied. This test was chosen due to small sample sizes (n = 6, months per season-year), non-normal data distribution, heterogeneity of variances, and the presence of outliers in some variables. The objective was to determine whether significant differences existed between seasons in recent years.

The results showed that Min_T, Max_T, Mean_T, Evap, Ther_O, and Helio during the dry seasons of 2011–2022 (a complete seasonal series) did not differ significantly from the median values of the rainy seasons during 2006–2010 (p > 0.05). This indicates that the current medians (2011–2022) of the dry season register statistically equal values to the medians that occurred in the rainy season of 2006 to 2010 (Table 8). Acum_P and Solar_R were characterized by showing significant differences between the climatic seasons.

Table 8 Tables of differences between groups year-season (2006–2022).
Full size table

Perkinsiella-climate relationship can be observed in Fig. 8, where Perkinsiella series (weather season-standardized and smoothed MA (3)) and temperature environmental parameters are compared.

Fig. 9
figure 9

Perkinsiella vs. and climatic parameter anomalies.

Full size image

Figure 9 shows how the peaks (amplitudes of positive and negative series) coincide between Perkinsiella series and atmospheric parameters. This suggests that during periods of negative temperature anomalies (i.e., temperatures below the monthly average), positive anomalies in Perkinsiella abundance (i.e., values above the monthly average) were also recorded (negative correlations in rainy season). A similar pattern was observed for other climatic parameters as well.

The results of this study demonstrate a progressive increase in temperature over time, which is associated with a higher incidence of Perkinsiella outbreaks. Furthermore, the data indicates that insect outbreaks not only occur during the dry season but also exhibit a marked tendency to persist throughout the year.

Discussion

In 2023, the values of Min_T, Max_T, and Mean_T from January to September recorded significantly higher positive anomalies compared to recent years. This is attributed to the high temperatures (Mean_T average of 27.0 °C) associated with the El Niño event and anthropogenic emissions. Additionally, there were locations where temperatures exceeded pre-industrial average temperatures by 1.5 °C on several days (during the summer solstice)5.

During this analysis period, the study area experienced ENSO events, and their effect on atmospheric parameters was evident through time series of anomalies, calculated relative to 1991–2020 period. The impact caused by the El Niño events of 2014–2016, 2018–2019 and 202336 generated considerable positive anomalies in all analyzed climatic parameters, particularly in temperature and Acum_P. Conversely, considerable amplitudes of negative anomalies, in the aforementioned variables, were observed due to the climatological phenomenon known as La Niña in the years 2007–2008, 2010–2012, 2016–2018, and 2020–202336. The period 2006–2023 did not record major ENSO events like those in the other periods, for example 1997–1998. However, in the period 2006–2023, the annual accumulated of Acum_P showed an increasing trend. This increase aligned with the expected behavior for Central and South America37.

The irregular seasonal occurrence pattern of Perkinsiella was also reported by Flores et al.38 at three study sites. Their findings indicated that infestations tipically begin during the dry season, but the highest population peaks occur during the rainy season (April and July in Valdez; November and March in Ecudos; and the first months of the rainy season at Ingenio San Carlos), demonstrating that the increase in temperature and relative humidity favor the development of the pest. Similarly Mendoza et al.39 observed significant fluctuations in population abundance throughout the year in Ecuador, noting that infestations generally begin during the dry season, persist through the rainy season, and peak in February or March. Moraga40 and several other authors10,12,13,15,16,41,42,43,44,45,46,47,48,49 agree that climatic conditions substantially influence insect embryonic and post-embryonic development, life cycle duration, population dynamics and density, geographical distribution, and survival. Consistent with these findings, Pulido Fonseca4 reported that environmental factors, especially temperature and relative humidity, affect the developmental duration of each life stage, in Hawaii, the life cycle of Perkinsiella lasted 48 days at 22 °C and 56 days at 25 °C, with the egg stage alone lasting 35–40 days at these temperatures.

In this study, the summer (cooler) season recorded an average temperature of 24.88 °C, while the winter (warmer) season averaged 26.54 °C. Minimum and maximum temperatures were 23.4 °C and 27.07 °C in summer, and 24.5 °C and 27.87 °C in winter, respectively—values consistent with those reported by Pulido Fonseca. Temperature increases, depending on the species, may influence developmental rates, the number of generations per year, population density, genetic composition, host plant use, and both local and geographic distribution. Moreover, elevated temperatures can favor pest outbreaks, weaken natural enemy populations, and reduce host plant resistance14. These alterations can be attributed to both natural global climate change and anthropogenic climate change, which have caused modifications in ecosystems and the extinction of species6.

In this regard, Mendoza et al.39 reported the appearance of the insect during the dry season and note temporary population increases during extended drought periods. Similarly, Contreras et al.50, in studies on the Central American locust (Schistocerca piceifrons Walker), observed population surges and concluded that the insect’s biological development cycle is triggered by rising temperatures and drought conditions. Wilson et al.51 documented that S. saccharivora (Delphacidae), a sugarcane pest in Louisiana, emerged as a significant threat during outbreak years—2012, 2016, 2017, and 2019—with high winter temperatures being associated with these events.

Climate change and extreme weather events exert profound impacts on crop production and pest dynamics. Insect pests, being highly adaptable, respond differently to various climate change drivers and increasingly pose a major threat to food security. Effective pest management strategies are essential to mitigate this issue. Therefore, understanding pest biology and behavior in relation to environmental changes is crucial, as climate change will impact their distribution and activity52,53.

Climate fluctuations, such as increasing temperatures and changes in precipitation, have an impact on agricultural production and insect pests. Global warming is associated with the geographic expansion of insects, increased winter survival, more reproductive cycles, and a higher risk of invasive species and insect-borne diseases. In tropical regions, insect populations may decline as temperatures approach or exceed their physiological optima, whereas increases in pest abundance are expected in temperate zones. Moreover, rising temperatures can indirectly impact crop yields by increasing atmospheric water demand, leading to water stress and reduced soil moisture availability. These conditions can also intensify the frequency and severity of heatwaves, which may influence pest populations (e.g., increased mortality of temperature-sensitive species), as well as the prevalence of weeds and plant diseases6,52,54,55,56,57,58,59.

Conclusions

The climatic conditions during the 2006–2023 period have shown a notable increase in anomalies, particularly in 2023, with significant deviations in temperature, precipitation, and solar radiation parameters. The weather in the study area exhibits two distinct seasonal patterns. The rainy season recorded the highest average minimum and maximum temperatures over the twelve months of each year. Temperatures showed notable variations; currently (2006–2023), there have been increases in means and medians primarily during the dry season (e.g., + 0.25 °C average and + 0.20 °C median in Mean_T).

Perkinsiella showed changes in its abundance from March to October since 2011, in both dry and rainy seasons. The medians of the temperatures during the dry season of recent years (2011–2022) are similar to those of the rainy season from a decade ago (2006–2010), indicating that temperatures are favorable for the insect, since it had only exhibited abundance during the rainy season in former years.

There is a clear relationship between Perkinsiella and the climate during the rainy season, where the highest absolute values (> 0.19) were recorded in Min_T, Helio, and Med_T. However, due to climate change, the values recorded in recent years during the dry season are similar to those of previous rainy seasons, creating a favorable scenario for insect reproduction throughout the years. These findings demonstrate the presence of climate change and its impact on biological elements such as Perkinsiella, which has increased its population in sugarcane shoots, leading to economic losses in production systems.