Spatial Distribution of Tibraca limbativentris (Heteroptera: Pentatomidae) in Babassou Palms

The babassou plant (Attalea speciosa Mart. Ex. Spreng, Arecaceae) is an important palm tree in the state of Maranhão, northeastern Brazil. This plant is the main arboreal component in extensive geographical areas including agricultural areas. However, limited research studies exist on its role as an alternative pest host in this region. This study investigated the occurrence of Tibraca limbativentris Stal (1860) in young babassou palm trees, the effect of abiotic factors (temperature, solar radiation and rainfall) on the number of T. limbativentris adults and postures, and this insect’s spatial arrangement during the rice crop off-season. The research team inspected young babassou plants on a monthly basis in 2012 and 2013 to collect and quantify postures and adults. They collected a total of 1418 live adults, 13 dead and ten postures of T. limbativentris. There was a significant difference between the first and the other collections of live adults carried out in 2012 and 2013. Abiotic factors including rainfall, temperature and solar radiation, had no influence on the number of adults (alive and dead) and postures. The semivariogram adjusted to the Gaussian model showed that in 2013, the spatial distribution of living adults, that had an aggregate pattern, was highly dependent on the season. However, in 2012 there was a pure nugget effect. These results imply that young babassou plants are alternative T. limbativentris hosts.


Introduction
The northern region of Maranhã o, northeastern Brazil, has extensive babassou palm trees (Attalea speciosa Mart. Ex. Spreng, Arecaceae). These trees constantly transform during rice crop implantation periods. The area's farmers practice a cultivation system that combines subsistence agriculture, artisanal fishing and extraction. This system uses fire to clean and fertilize the soil by leaving the residual ash from burning the fallow bushes on the ground. Such techniques have made the state of Maranhã o the largest rice producer in the northeast Brazilian region. Furthermore, this state ranks second among states with the most extensive rice plantations in the country (CONAB, 2019).
The phytophagous bugs are pests of economic importance to rice farmers. This fact is mainly true for the thatched bugs, the hornbills, Tibraca limbativentris Stal, 1860 (Heteroptera: Pentatomidae) that are the main pests in South America (Panizzi, 2015). This pentatomid contributes to up to 90% of farmer's productivity losses (Souza et al., 2008). The pest's adults and nymphs feed on the plant's vegetative and reproductive phases resulting in the symptoms of a dead heart and a white panicle, respectively (Botton et al., 1996). Therefore, the bug causes the most damage in the rice crop's reproductive phase (Krinski & Foerster, 2017). The bug is also known as "cangapara" in the state of Maranhã o. This bug is found in almost all the state's municipalities that grow upland rice.
Due to these bugs' polyphagy, they seek alternative plant hosts to feed, oviposition and develop their offspring (Panizzi, 1991); thus, they hibernate in adjacent crop remains or alternative host plants waiting for the next rice crop. A behavior confirmed by Pasini et al. (2018) who demonstrated the root bug in clumps of foxtail grass, Andropogon bicornis L and A. lateralis (Poaceae) in the rice crop's off-season period. According to Engel et al. (2019), these alternative resource sources help bugs survive unfavorable periods.
Because they are highly resistant to fire, babassou plants are the dominant species in Maranhão's cultivation areas after a rice harvest (Mitja et al., 2018). Therefore, the babassou palm may be a potential shelter for T. limbativentris in the off-rice season.
Advances in integrated pest management have permitted study of the spatial distribution of some bug species in agricultural areas to construct infestation maps for targeted and precise application of population reduction measures. For instance, the spatial distribution of T. limbativentris adults and nymphs in irrigated rice fields depends on the crop's phenological stage (Costa et al., 2019;Pasini et al., 2020;Pasini et al., 2021). Engel et al. (2021), point out that the spatial distribution study in alternative hosts helps in pest permanent monitoring and anticipating possible outbreaks.
Despite the potential economic damage caused by bugs, scarce information exists on its spatial distribution in alternative hosts. Therefore, this research aimed to describe the distribution of T. limbativentris in young babassou plants adjacent to rice growing areas in the northern region of the state of Maranhã o. This study also investigated the influence of abiotic factors (temperature, solar radiation and rainfall) on the number of the pests' adult populations and spawnings and the insect's spatial distribution in the area inter-harvest period.

Study Area: Location and Characteristics
The research team conducted a study in the municipality of Matões do Norte (latitude 03º 3757´ S; longitude 44º 339´ W), (05º 42´ S; 43º 13´ W), State of Maranhã o, northeastern Brazil. They did the research in areas where family farmers cultivated upland rice intercropped with corn (Zea mays L.) and cassava (Manihot esculenta Crantz).

Climatic Data
The study team obtained the average 30 days' temperature (ºC), solar radiation (cal cm -2 day) and rainfall (mm) from the Meteorological Laboratory of the Geoenvironmental Nucleus of the Maranhã o State University (Figure 1).

Sampling of T. limbativentris
The team collected samples each month after the rice harvest from August to December 2012 and August to October 2013. Each year, they selected an area measuring approximately four hectares surrounded by young babassou palms as their study sample. The areas' predominant cultivation system was "slash and burn", where vegetation was cut, burnt to clean the site and the ash used as a soil fertilizer. This soil preparation generally occurred in November, followed by rice sowing in January to coincide with the beginning of the region's rainy season. Nevertheless, the areas were kept fallow for vegetation regeneration after a rice harvest.
The selected young babassou plants (called pindovas), measuring 60 cm in height and 52 cm in diameter, were the most vigorous plants located equidistantly (approximately 5m apart) within the rice planting area.
The team randomly selected pindovas' sampling points following a zigzag pattern covering the entire area related to the rice clump studied.
The team marked 64 and 49 pindovas in 2012 and 49 in 2013. They meticulously inspected pindovas leaves and trunk for the presence and collection of live and dead adult bugs, and T. limbativentris postures

Georeferencing of Sampled Plants
The study used a Garmin MAP 76CSX GPS (Global Positioning System) device to georeference the selected babassou plants for T. limbativentris sampling. The device used the flat coordinates of the Universal Transverse Mercator System (UTM) to identify the palm trees' location and facilitate repeated sampling from the same plants.

Statistical Analysis
First, the Shapiro-Wilks normality test was used to describe data collected on adult bugs (living and dead) and postures. Then, since data was non-normal, the Kruskal-Wallis non-parametric test, p <0.05 (Statsoft Inc, 2011), was used to identify statistical differences between the number of adults (living and dead) and collected postures.
Subsequently, the number of live and dead adults and postures were transformed into their square roots (√x) and tested using the T-test at the level of 5% probability in the BioEstat 5.0 program (Ayres et al., 2007). Finally, pearson's linear correlation was used to describe the relationship between abiotic factors (temperature, solar radiation and rainfall) and the number of adults (living and dead) and postures.

Geostatistical Analysis
A Geostatistical analysis was used to construct semivariograms and verify spatial dependence between samples. According to Cambardella et al. (1994), the [(Co / C1 + Co) x100] model's degree of spatial dependence is considered strong if this ratio is ≤ 25%, moderate if it is between 25 and 75%, and weak when the ratio is > 75%. Subsequently, the variograms were constructed using the model with the best determination coefficient.

Preparation of Spatial Distribution Maps
The study developed spatial distribution maps using all the annual data on live adult bugs. The data obtained were imported into SAS version 8.2 (SAS Institute, 2001) for analysis then, crossed and interpolated. In addition, bug infestation maps were generated using the G.S. + program (Gamma Design Software, 2000).

Results and Discussion
A total of 1418 live adults, 13 dead and 10 postures of T. limbativentris were collected in 2012 and 2013. The high number of live individuals collected from the pindovas suggests the presence of a surviving T. limbativentris population large enough to infest the same rice fields the following year. Bugs infest rice plants within 20 or more days of emergence and begin oviposition ten days later (Ferreira, 1998), with the possibility of reproducing up to three  Therefore, it is possible to infer that young babassou plants present microclimatic conditions that favor the survival of the T. limbativentris population. According to Klein et al. (2013), a protected environment is a better shelter since it provides more stable microclimates than the open environments.
The rainfall, temperature and solar radiation did not influence the number of adults (alive and dead) and postures collected between August and December 2012, and August and October 2013 ( Table 1).
The absence of a correlation or presence of low correlation between climatic factors and pest populations may indicate that these abiotic factors did not affect the area's T. limbativentris population. However, the number of T. limbativentris adults and postures in each collection was insufficient to illustrate the influence of abiotic factors on this insect-pest's population. ISSN 2166-0379 2021  Based on the above parameters, nugget effect, level and range for the period correspondent to the year 2012, there was a random pest distribution in the babassou young plants (Table 2). Thus, the appropriate model to describe the spatial behavior of adults of T. limbativentris was a pure nugget effect ( Figure 3A).

Journal of Agricultural Studies
A different research study conducted in irrigated rice fields substantiated the same distribution of adults and nymphs as T. limbativentris (Alves et al., 2016). This pure nugget effect in insect collections is due to sampling error or spatial dependence; however, this usally occurs on a smaller scale than the one we adopted (Liebhold et al., 1993). Thus, there was a random distribution of live adult bugs; alternatively, or that the distance between the sampling points in the area was too great to identify a dependent spatial distribution. This distance could be attributed to the disposition of the young babassou plants in the area. ISSN 2166-0379 2021  The distribution of bugs in 2013 was adjusted to the Gaussian model ( Figure 3B), due to the great continuity of the variables, that is, values found very close to semivariance (Table 2). Similarly, Pazini et al. (2015) found that this model was the best that was adjusted in most T. limbativentris collections in irrigated rice.

Journal of Agricultural Studies
In 2013, the semivariance initially increased as the distance between the sampling points increased until a point where the semivariance reached a plateau ( Figure 3B). However, the semivariance for the year 2012 remained constant throughout the sampling distances ( Figure  3A). As for the adjustment's efficiency, the Gaussian model obtained an R 2 = 0.67; this figure indicates a good adjustment based on Silva et al. (2015). The number of adult live wire bugs collected in 2013 had a range of 29.45 m. This parameter represents the limit of dependence on the sampled points.
Based on the criterion by Cambardella et al. (1994), the 2013 samples portrayed strong spatial dependence, because it was ≤ 25%, indicating that the number of individuals depends on their position in the area. The kriging map was generated for the year 2013 only. This map shows that the spatial distribution of T. limbativentris was aggregated, with a concentration of specimens in the young babassou plants located at the area's edge ( Figure 4). When assessing the spatial distribution of Helicoverpa armigera (Hubner) (Lepidoptera: Noctuidae) and Pectinophora gossypiella (Saunders) (Lepidoptera: Gelechiidae), Milonas et al. (2016) reported that these pests' spatial distribution depends on the area's availability and distribution of hosts. Following this study, Martins et al. (2020) mentioned that it is possible to expand the study of insect behavior in the field from their study of the spatial distribution of Neomegalotomus parvus (Hemiptera: Alydidae). Young babassou plants are widely distributed on the edges and within rice-growing areas. Consequently, these plants play an essential role in T. limbativentris survival during the rice crop off-season. Therefore, program managers can focus appropriate strategies on young babassou plants to manage and control T. limbativentris to reduce this pest's incidence in the new rice planning season.

Conclusions
The young babassou plants serve as a refuge for Tibraca limbativentris in the rice crop off-season. This shelter implies that abiotic factors, such as temperature, humidity and precipitation, have no impact on the study area's insect population. Furthermore, aggregates of fully-grown T. limbativentris are spatially distributed in this alternative host and form ridges at the area's edge. Such a location favors sampling plans for the management and control of this pest that economically influences rice crop production.