Rice Responses to the Stem Borer Diatraea saccharalis (Lepidoptera: Crambidae) by Infrared-Thermal Imaging: Implications for Field Management

Abstract

Diatraea saccharalis (Fabricius) is a major pest of rice crops, and its early detection—before any visible plant damage occurs—is crucial to prevent yield losses and establish effective, rational control methods. This study aimed to model the infrared-thermal responses of rice cultivars to D. saccharalis infestation levels. Between 2019 and 2020, two experiments were conducted in a controlled environment using the cultivars IR 40 and BR IRGA 409, previously identified as having different resistance reactions. Rice plants grown in pots were manually infested with first-instar larvae of D. saccharalis, ranging from 0 to 10 caterpillars per plant, with the plants maintained in cages covered with voile fabric throughout the trial. After 30 days of infestation, the number of live and dead caterpillars, the number of damaged and healthy stems, and the dry mass of the aerial parts were evaluated. A generalized linear mixed model was applied to the data obtained from leaf temperature as a function of infestation level throughout the infestation period, using the F-test to detect significant differences between cultivars. Generalized Additive Models for Location, Scale, and Shape (GAMLSS) were fitted to the variables related to resistance. It was observed that leaf surface temperature is related to the level of infestation and could be used to detect susceptibility in IR 40. In both cultivars, leaf temperatures were higher within the first 15 days post-infestation.

1. Introduction

Diatraea saccharalis Fabricius (Lepidoptera: Crambidae) is an agricultural pest that inhabits the Americas, mainly attacking sugarcane (Saccharum spp.), corn (Zea mays L.), sorghum (Sorghum bicolor L. Moench), and rice (Oryza sativa L.). The damage caused by D. saccharalis to these crops is economically significant [1,2,3,4], and results from the drilling and opening of galleries along the stems by the caterpillars, destroying the vascular system [5]. As the caterpillars penetrate and develop inside the stems, it is possible to observe a mass composed of plant tissue and fecal residues close to the entrance closure [2].

Visual detection of the symptoms, the infestation, and, above all, its potential for damage is quite difficult. In rice, D. saccharalis has been difficult to control, because when the main symptoms that demonstrate its attack, “dead heart” and “white panicle”, are noticed, the damage inside the stems has already occurred, making it very difficult to exterminate the caterpillars inside. Another important aspect is that the number of attacked stems is always higher than the number of stems showing both symptoms [2]. For this reason, the recommendation to monitor the insect during the stages of greatest rice susceptibility has been reinforced, around 50 days after seedling emergence, by recording the presence of eggs and adults.

However, using traditional sampling methods, the practice of monitoring this pest becomes highly expensive and challenging for tropical and long-scale agriculture, such as in Brazil. The pest’s rapid and unpredictable population growth requires frequent surveys, for if spaced just a week apart, they can initially show a balanced situation and, later, populations that are found inside the stems at densities that exceeded the action levels. In this case, early detection is crucial to avoid yield loss, as effective treatments at the correct time can allow plants to recover.

The development of advanced electronics, Geographic Information Systems, and remote sensing have enabled significant advances in the practice of precision agriculture. Irrigation, fertilization, disease and weed detection, and yield mapping are some examples of crop management practices that are being transformed by remote sensing [6]. Recent technological advances in imaging and sensor networks have demonstrated the technique’s potential for application in automated monitoring of arthropod pests, optimizing the use of insecticides, and reducing yield losses in large-scale agriculture [7,8].

Among the remote sensing techniques, thermography can potentially be used in agriculture in the areas of plant physiology and plant protection. In recent years, a broad spectrum of applications has been found for thermal infrared imaging in plants, based on its relationship to transpiration and stomatal conductance. Thermal infrared remote sensing detects energy emitted in the thermal infrared band (8–14 μm), which is reflected by vegetation over long distances, and can be used to detect changes in plant transpiration and water content caused by biotic and abiotic factors, as well as to identify changes in the temperature of the plant canopy as a result of changes in physiological variables related to stress [9]. For example, infrared thermal analysis has been used to (i) evaluate plant responses to soil water availability conditions [10,11], (ii) high-throughput phenotyping and selection of drought tolerant genotypes [12,13,14], (iii) estimation of chlorophyll concentration [15], (iv) injury detection and/or presence in cultivated plants or stored grains caused by different biological agents, such as phytopathogens [16,17,18,19,20,21,22], and (v) pest arthropods [23,24,25].

With the hypothesis that leaf temperature can be used to discriminate rice cultivars in relation to susceptibility and damage caused by infestation of D. saccharalis, the objective of the present study was to evaluate and model the responses of two rice cultivars, IR 40 and BR IRGA 409, of contrasting susceptibility to D. saccharalis attack using infrared thermal imaging.

2. Results

2.1. Insect Damage and Plant Variables

Infestation by D. saccharalis resulted in an increase in leaf temperature, with a nearly linear temperature gradient observed from 0 to 10 caterpillars per plant (Figure 1).

The proportion of live caterpillars increased with the level of infestation, showing no significant difference between cultivars at any infestation level (p > 0.05). By the end of the experiment, fewer than 40% of caterpillars remained alive in both cultivars (Figure 2).

The proportion of healthy stems decreased linearly as the number of D. saccharalis caterpillars per plant increased in both cultivars, indicating significant differences (p < 0.05) at all infestation levels (Figure 2). Overall, IR 40 exhibited more damaged stems than BR IRGA 409. Specifically, the proportion of healthy stems in IR 40 decreased from 0.69 to 0.53 with increasing infestation, while BR IRGA 409’s healthy stems decreased from 0.86 to 0.72.

Plant dry matter did not show a clear trend as infestation levels rose, but BR IRGA 409 generally had higher values.

2.2. Thermometric Measurements

Significant differences (p < 0.05) were observed in the average leaf surface temperatures of the rice cultivars throughout the experimental period, across all levels of infestation (Figure 3).

Although non-infested plants of IR 40 exhibited higher leaf temperatures, both cultivars responded to the level of infestation. While BR IRGA 409 showed a slightly more sensitive response to infestation levels, as indicated by the orientation of the regression line, IR 40 maintained higher average leaf temperatures, ranging from 26 to 28.5 °C, whereas IR 40 exceeded 30 °C.

Both cultivars exhibited higher leaf temperatures during the first 15 days after infestation, particularly in IR 40, specifically between the 50th and 65th days after sowing (Figure 3). The temperature differences between cultivars were further highlighted by the calculated difference between air and leaf surface temperatures throughout the infestation period. IR 40 consistently showed a leaf surface temperature closer to the air temperature.

3. Discussion

Infrared thermal imaging has emerged as a tool for detecting pest insects in cultivated plants due to its non-invasive and rapid assessment capabilities, based on subtle temperature variations caused by pest infestations, allowing for timely interventions that can significantly reduce damages. Economically, especially for agricultural crops, the early identification of pest threats translates into lower pesticide volume and reduced crop loss.

Leaf surface temperature values were significantly higher in the IR 40 cultivar, indicating a relationship with biotic stress. Based on the results from experiments conducted over two agricultural years, this study found that thermal infrared images were effective in differentially diagnosing D. saccharalis injuries in rice plants. IR 40 displayed a susceptibility response, while BR IRGA 409 demonstrated a resistance response [5].

Rapid and accurate quantification of pest attack symptoms is crucial for integrated pest management (IPM) and crop protection, particularly for preventing early economic damage. This approach is essential for reducing insecticide use, thereby enhancing both profitability and environmental sustainability [26].

The injury caused by D. saccharalis feeding on the inner part of rice stems affects plant physiology similarly to water stress. As demonstrated by Godfrey et al. [27], plants damaged by gallery formation from stem borers or experiencing water stress typically show a reduction in net photosynthetic rate, stomatal conductance, and intercellular CO₂ concentration, along with an increase in leaf temperature [24]. It is likely that feeding on the inner rice stalks disrupts water movement from the soil to the upper photosynthetic leaves. Consequently, the plant struggles to transport water through the vascular system, which is obstructed or destroyed by the caterpillars, even when the roots are in adequately moist soil. This leads the plant to respond as if it were under water stress, partially closing its stomata to conserve water and reduce transpiration loss.

The cultivar IR 40, having the highest proportion of damaged stems, exhibits an increase in leaf surface temperature, explaining the observed differences compared to BR IRGA 409. Soroker et al. [28] demonstrated that a thermal monitoring system could detect palm trees infested with larvae of the red palm weevil, Rhynchophorus ferrugineus (Olivier) (Coleoptera: Curculionidae), approximately three weeks before visual symptoms appeared. According to the authors, larval attacks induced water stress, resulting in higher temperatures in the plant canopy compared to uninfested palm trees. The potential of thermal infrared remote sensing techniques for the early detection of pests in crop plants has been supported by several studies [29,30,31].

The relationship between arthropod population levels or damage to cultivated plants and production loss can be directly influenced by resistant cultivars, each distinguished by their resistance category [32,33]. In this study, thermal images effectively detected D. saccharalis infestation in both cultivars, regardless of the level or period of infestation. However, leaf temperatures were significantly higher in the IR 40 cultivar compared to the resistant BR IRGA 409 [5]. Cultivars that show smaller changes in physiological parameters—such as photosynthetic rate, transpiration, and stomatal conductance—when exposed to insect attack typically exhibit greater tolerance to pests [34]. Therefore, the development and validation of automated monitoring systems and technologies, such as thermal imaging, must consider the cultivar’s response to pests to avoid underestimating or overestimating the results.

Our study demonstrated the potential of infrared thermal imaging for detecting D. saccharalis infestation and associated symptoms in rice cultivars. This represents an initial evaluation of a non-invasive detection method for D. saccharalis before its widespread establishment. Although the prospects are encouraging, given that stem borer sampling is both destructive and labor-intensive in rice fields, further research is needed to adapt this technique for field application. Its high sensitivity to environmental factors, such as temperature, radiation, rain, and wind, can significantly influence estimates [17].

Other types of plant stress, such as water and nutrient deficiencies, can lead to variations in leaf temperature, complicating the interpretation of infrared thermal imaging data, which can obscure accurate detection of pest infestations by infrared thermal imaging. Future research could focus on elucidating the specific physiological mechanisms behind temperature variations and developing models that can differentiate between stressors. Future studies should consider analyzing plant responses under multiple stressors [8]. For example, investigating simultaneous attacks by different pest species or the effects of pest infestations on plants with and without water or nutritional deficiencies could provide valuable insights.

4. Materials and Methods

4.1. Study Site and Plant Material

Two experiments were carried out in two agricultural years, harvests 2019/2020 and 2021/2022, both at the Instituto Federal Goiano, Campus Urutaí, located in the Brazilian Cerrado, State of Goiás (17°27′49″ S; 48°12′06″ W; average altitude of 807 m). Two commercial rice cultivars, IR 40 and BR IRGA 409, were selected for the experiments. This selection was based on a broad study on evaluating the resistance of rice accessions to D. saccharalis [5], in which IR 40 and BR IRGA 409 demonstrated susceptibility and resistance behavior, respectively. The cultivar IR 40 has “modern type” plants, a cycle of 119 days, and a productive potential of 5.0 t ha⁻¹ [35]. BR IRGA 409 is a cultivar that also presents “modern type” plants, cycle of 120–130 days, and productive potential of 10.1 t ha⁻¹ [36].

The seeds were sown in plastic trays with a commercial substrate (Maxfertil, Pouso Redondo, SC, Brazil). Twenty days after emergence, three seedlings were transplanted into 5 L plastic pots filled with Red Oxisol, with 42% clay, taken from an experimental area with a known use history, receiving 0.75 kg of limestone per m³ of soil and 200 g of the formulated fertilizer 08-30-10 (NPK) per m³ of soil. The pots were watered daily by the greenhouse’s irrigation system and remained in this environment throughout the experiment period. Cultural treatments were carried out in accordance with the technical research recommendations for rice cultivation, but without the application of insecticides.

4.2. Insects and Experimental Procedures of Infestation

D. saccharalis eggs were obtained from Laboratório Biocana, located in Itumbiara, Goiás, Brazil. Egg masses were kept in plastic cups (300 mL) in a controlled environment (25 ± 2 °C; 12:12 h (Light:Darkness); 70% relative humidity) until the caterpillars hatched.

Fifty days after sowing, plants were infested with first-instar caterpillars of D. saccharalis. To do so, Eppendorf-type tubes containing caterpillars were distributed over the plant sheaths in each pot, which were covered with a voile fabric cage (0.5 × 0.5 × 1.0 m) throughout the experiment to avoid insect escape and natural enemies. For both cultivars, the following infestation levels were used: 0, 2, 4, 6, and 10 caterpillars/plant. Each infestation level was replicated on three plants. A completely randomized experimental design was adopted, consisting of five treatments (infestation levels) and three replications, totaling 15 cages or plots for each cultivar (IR 40 and BR IRGA 409).

4.3. Data Collection and Analysis

At three-day intervals after manual infestation, digital images of the experimental units were captured using an FLIR^® C2 IR camera (FLIR Systems, Wilsonville, OR, USA), with a 41° × 31° field of view, 60 × 80 pixel of optic resolution, spectral range of 7.5–14 μm, radiometric resolution of 0.1 °C, absolute precision of 2 °C, and emissivity adjustment ε = 0.98. The images were taken together with a digital thermo-hygrometer model TTH 100 (Incoterm, Brasília, Brazil) to obtain the atmospheric temperature and relative humidity, which were recorded between 10 a.m. and 2 p.m., at a standard height of 1 m above ground level (top of the cages). Leaf surface temperature was obtained from three random points on the plant (Figure 1).

After 30 days of infestation, at the end of the experiment, the plants were cut off at the base of the stems and taken to the laboratory to evaluate the number of live and dead caterpillars (NLC, NDC), number of damaged and healthy stems (NDS, NHS), and aerial part dry mass (DM, g) in an oven (105 ± 2 °C for 24 h).

A generalized linear mixed model was fitted to data obtained from leaf temperature as a function of infestation level throughout the infestation period (repeated measures), using the F-test to detect significant differences (p < 0.05) between cultivars. Generalized Additive Models for Location, Scale, and Shape (GAMLSS) [37] were fitted to data of the variables related to resistance: plant dry matter (normal distribution) and the proportions of NLC and NHS (beta distribution). Akaike’s information criterion (AIC) and the root-mean-squared-error (RMSE) were used to choose and evaluate the goodness-of-fit of models. The analyses were performed with the R software v.4.1.1 (www.r-project.org (accessed on 12 August 2021)).