Comparison of GARP and MaxEnt in Modeling Current and Future Geographic Distribution of Ceracris nigricornis Walker (Acrididae, Orthoptera) in China
Funding: This work was supported by Opening Foundation of Key Laboratory of Sichuan Province for Bamboo Pests Control and Resource Development, ZLKF202305, ZLKF202202, Science and Technology Program of Leshan Normal University, 2022SSDJ005, KYPY2023-0006, XJR17005, LZD010.
ABSTRACT
Ceracris nigricornis Walker is an insect of the Acrididae, which can harm bamboo, rice, corn, sorghum and other crops, and can cause serious economic losses. In this study, based on 234 occurrence sites of C. nigricornis obtained from the Global Biodiversity Information Facility and literature, and data of three future climate scenarios presented by CMIP6, two niche models (GARP, MaxEnt) were used to predict the suitable area of C. nigricornis in China. The result shows that the main environmental factors affecting the distribution of C. nigricornis are min temperature of coldest month (bio6), mean temperature of coldest quarter (bio11), precipitation of driest month (bio14) and precipitation of wettest quarter (bio16). From the result of MaxEnt model, it can be seen that the suitable area of C. nigricornis in China is 128.91 × 104 km2 under current scenario. It will decrease by 3.19% in the 2050s and then increase by 12.04% in the 2090s under the SSP1-2.6 scenario, increase by 5.79% in the 2050s and then decrease by 7.53% in the 2090s under the SSP2-4.5 scenario, and increase by 33.03% in 2050s and then decrease by 23.31% in the 2090s under SSP5-8.5 scenario. From the result of GARP model, it can be seen that the suitable area of C. nigricornis in China is 166.09 × 104 km2 under current scenario. It will increase by 8.41% in 2050s and then continue to increase by 6.11% in 2090s under SSP1-2.6 scenario, increase by 23.84% in the 2050s and then decrease by 0.88% in the 2090s under the SSP2-4.5 scenario, and increase by 34.37% in 2050s and then decrease by 1.75% in 2090s under SSP5-8.5 scenario. The boundaries of suitable areas will expand to the north and southwest of China under future climate change scenarios, specially in Sichuan, Chongqing and Yunnan. Local forestry authorities should strengthen the monitoring of bamboo forests to prevent the damage caused by the introduction of C. nigricornis.
1 Introduction
Ceracris nigricornis Walker (Orthopteridae: Acrididae) is a severe grasshopper pest of bamboos and is mainly distributed in China, Korea, India, Myanmar, Laos, Vietnam, Nepal and Thailand. There is no literature on whether C. nigricornis is a native or exotic species, but some studies suggest that its transmission route in China may start in Yunnan and spread through Guangxi, Hunan, Sichuan, Hubei and Anhui to Shaanxi (Wang 2019). C. nigricornis can harm bamboo, rice, corn, sorghum and other crops and can cause serious economic losses (Yuan et al. 2019). The adult and nymphs of C. nigricornis feed on the leaves of plant, biting the bamboo leaves into pure toothed-like grooves and even eating them up. As it can be seen that C. nigricornis have become very common in southern China and are very harmful to crops, it is very necessary to study them. Many scholars have studied its biological characteristics (Feng et al. 2024; Gao 2017), but no reports of its disaster risk assessment and regionalization in China have been found, which is an important prerequisite and basis for effectively controlling it.
Niche model, contains two types: time-based and space-based, is a mathematical model to study the relationship between various organisms in an ecosystem and their adaptability to the environment. The time-based niche model generally takes the dynamic change of biological population as the research object, whereas the space-based niche model mainly studies the space occupation between biological communities. The space-based niche model include Bioclim (Busby 1991), Domain (Carpenter et al. 1993), GARP (Stockwell and Peters 1999), ENFA (Hirzel et al. 2002), and MaxEnt (Phillips, Anderson, and Schapire 2006; Phillips and Dudik 2008), which the most commonly used are MaxEnt and GARP. MaxEnt model uses the distribution points of species and environmental variables to calculate the ecological niche of species and simulate the potential distribution of species (Xu, Peng, and Peng 2015). It has high prediction accuracy and ease of use (Merow, Smith, and Silander 2013), and has been widely used in the prediction of species potential distribution, the risk assessment of alien species invasion, and the study of the impact of climate change on biodiversity. Combining the optimization ability of genetic algorithm with the environmental interpretation ability of niche model, GARP model provides a powerful tool for predicting species distribution. It can be adapted to various data types and research problems and has been widely used to predict the species distribution of different taxa and pathogens in landscapes since the late 1990s (Haase et al. 2021). Based on the above advantages, MaxEnt model and GARP model as the tools of this study can better predict the potential distribution of C. nigricornis.
As a big agricultural country, China is very sensitive to climate change. Climate change not only directly affects the ecological habits of locusts but also indirectly affects the distribution and outbreak of locusts through other environmental factors (Liu, Zhang, and He 2024), which will seriously affect agricultural production and food security. The Climate Model Intercomparison Project (CMIP) is an international collaborative project initiated by the World Climate Research Programme (WCRP), whose aim is to understand past, present and future climate change by collecting and comparing simulation results from various global climate models (GCMs) (Zhou, Zou, and Chen 2019). At present, the 33 research teams involved in CMIP6 have registered 112 versions of the global climate system model. For the southwestern China, the ACCESS−CM2, CMCC−CM2−SR5 and CMCC−ESM2 global climate mode have the best performance in the comprehensive simulation ability of temperature (Jin, Jiang, and Zhang 2022).
In this paper, the Jackknife test method was used to analyze the key environmental factors affecting the survival of C. nigricornis and two niche model (MaxEnt and GARP) combined with ArcGIS were used to simulate the potential suitable areas of C. nigricornis in China. The purpose of this study is to answer the following questions: What are the key environmental factors affecting the distribution of C. nigricornis? What is the relationship between key environmental variables and the probability of the presence of C. nigricornis? What are the potential suitable areas of the C. nigricornis in China under current climate scenario? and How will climate change in future affecting the potential suitable areas of the C. nigricornis? The answers to the above questions is of great significance to research formulate feasible countermeasures to prevent and control C. nigricornis, reduce the serious economic loss caused by it, and ensure the stable development of bamboo and crop industry in China.
2 Materials and Methods
2.1 Data Sources and Processing
The distribution records of C. nigricornis were derived from the Global Biodiversity Information Facility (GBIF 2023) and literature (Gao, Jiang, and Fan 2020; Wu 2005; He 2002). Google Earth was used to retrieve the records missing the latitude and longitude, and ENMTOOL software was used to remove duplicate records in 2.5 arc-min grid (Zhu, Liu, and Gao 2014; Sillero and Barbosa 2021). After the abovementioned treatment, 234 occurrence records were reserved for species distribution prediction (Table S1). The specific points are in Figure 1.
In this study, we chose the ACCESS−CM2 model developed by commonwealth scientific and industrial research organization (CSIRO) for future climate simulations. We chose SSP1-2.6 scenario, SSP2-4.5 scenario and SSP5-8.5 scenario to represent different future climate change, of which the SSP1-2.6 scenario represents global carbon dioxide emissions will significantly reduce in the future, reaching net zero emissions by 2050. The SSP2-4.5 scenario represents global carbon dioxide emissions will hover around current levels before starting to decline by mid-century, reaching net zero by 2100. The SSP5-8.5 scenario represents global carbon dioxide emissions will rise steadily in the future, reaching about double current levels by 2100.
The environmental factors data contain 19 bioclimatic factors and altitude, which can be downloaded from WorldClim (https://www.worldclim.org/). The file format of the environmental factor is tif with a spatial resolution of 2.5 arc-min grids, and ArcGIS 10.0 was used to convert it to “ASC” format for future use. Contribution rate of environmental factors to predictive model construction was tested by the jackknife method, and the factors with zero contribution rate were deleted. To reduce the impact of multiple contributions of environmental factors on the model and avoid errors in modeling caused by multicollinearity, SPSS 14.0 software was used to conduct Pearson test, and the one with the lower contribution percentage value in the jackknife test among the two factors with correlation coefficients > 0.85 was eliminated. Finally, the variables and their factors used for species modeling are shown in Table 1.
| Code | Environmental factor | Unit |
|---|---|---|
| Bio3 | Isothermality | — |
| Bio5 | Max temperature of warmest month | °C |
| Bio6 | Min temperature of coldest month | °C |
| Bio11 | Mean temperature of coldest quarter | °C |
| Bio14 | Precipitation of driest month | mm |
| Bio16 | Precipitation of wettest quarter | mm |
| Alt | Elevation | m |
2.2 Niche Models and Parameter Settings
The software containing the MaxEnt model was downloaded from the official website (V3.4.4, https://biodiversityinformatics.amnh.org/open_source/maxent/). The geographical distribution records of C. nigricornis in “CSV” format was imported into the “Samples” data boxes of MaxEnt software, whereas the environmental factors in “ASC” format was imported into the “environmental layers” data boxes. The options on the right side of the software, such as “Create response curves” and “Do jackknife to measure variable importance”, were selected. The feature combinations (FCs) option and regularization multiplier (RM) option were used to optimize the accuracy of the MaxEnt model. The FCs option contains linear (L), quadratic (Q), hinge (H), product (P) and threshold (T), which can be formed as L, LQ, H, LQH, LQHP and LQHPT. The RM step was set to 0.5 and ranges from 0.5 to 4. Therefore, 48 different combinations of RM and FCs were established and the combinational model with the minimum Akaike information criterion correction (AICc) value is determined to be the optimal model (Guo et al. 2022). “Random test percentage” was set to 25% in the initial model, whereas the “replicates” was set to 10 in the reconstructed model. The software containing the GARP model was also downloaded from the official website (V1.1.6, http://www.lifemapper.org). The training data and the test data set were 75% and 25% of the distribution data, and the number of runs was set to 20. Beyond that, all other basic settings are set to default values.
2.3 Model Accuracy Evaluation Method
The ROC curve (Receiver Operating Characteristic curve) is a graphical tool used to evaluate the performance of binary classification models, which demonstrates the performance of the model by plotting True Positive Rate (TPR) and False Positive Rate (FPR) under different classification thresholds. The closer the ROC curve is to the upper left corner, the better the performance of the model, that is, the lower the false positive rate and the higher the correct recognition rate. The area under ROC curve (AUC), the area formed by the ROC curve and the axis, is another important indicator to measure the model performance. Many times the ROC curve does not clearly indicate which classifier performs better, but AUC as a value can perform better. Therefore, we use the AUC value as a criterion for evaluating the model for its discrimination capacity (Lobo, Jiménez-Valverde, and Real 2008). The closer the value of AUC is to 1, the better the model performance (Xu 2024). If the AUC value is > 0.9, the prediction effect of the model is better; If the AUC value is < 0.9 and > 0.7, the prediction effect of the model is general; If the AUC value is < 0.7, the prediction effect of the model is poor (Merow, Smith, and Silander 2013; Wang et al. 2007; Wang, Zhu, and Xu 2021).
2.4 Suitable Area Division
The calculation results in Ascii format of MaxEnt and GARP were loaded in ArcGIS 10.2, and the suitable biological grade was divided and visualized to obtain the potential distribution map of species (The projection coordinate system was GCS_WGS_1984). It is important to select the appropriate threshold value when converting the results of successive species suitability predictions into the boolean form of “suitable area” and “unsuitable area”. A large number of studies have proved that the threshold method of sensitivity-specificity and maximization approach is significantly superior to other threshold methods (Jiménez-Valverde and Lobo 2007), and the threshold of the existence probability of a specificity is 0.33. The region with the occurrence probability of species < 0.05 was defined as the unsuitable area of C. nigricornis, the region with the occurrence probability of 0.05–0.33 was defined as the lowly-suitable area, the region with the occurrence probability of 0.33–0.66 was defined as the moderately-suitable area, and the region with the occurrence probability of 0.66–1.00 was defined as the highly-suitable area (Sun, Qin, and Liu 2012). The changes of the potential distribution area of C. nigricornis were mainly calculated using the SDM toolbox (Etherington 2011), which was mainly written based on python. We calculated the percentage change of suitable area from current to the 2050s and from the 2050s to the 2090s.
3 Results
3.1 Model Accuracy Evaluation
According to the known occurrence data and current climate data, the potential geographical distribution map of C. nigricornis was modeled and the region climatically suitable was determined. The AUC values of two niche models (MaxEnt and GARP) for 10 replicates were all > 0.92 (Table S2), indicating that the performance of the distribution model was better than that of the random model, and the stability between each repetition was good. The model had excellent performance in predicting the suitable area of C. nigricornis.
3.2 Key Environmental Variables Affecting the Geographical Pattern
The jackknife method of Maxent software was used to test the importance of seven environment variables to modeling (Figure 2 and Table 2). The results show that the regularization training gain of min temperature of coldest month (Bio6) is 1.55, the percent contribution is 18.3%, and the permutation importance is 39.7%, all of which are high levels. It can be concluded that min temperature of coldest month is the most important variable affecting the distribution of C. nigricornis. The regularization training gain value of precipitation of wettest quarter (bio16) is 1.42, and the contribution rate is 44.5%, indicating that this variable is also very important. Similarly, mean temperature of coldest quarter (bio11) and precipitation of driest month (Bio14) are also important factors affecting the distribution of C. nigricornis. In the process of variable elimination modeling, the regularized training gain value of isothermality (bio3) and max temperature of warmest month (bio5) decreased the most, indicating that this variable contains important information with the C. nigricornis distribution. According to the above comparative analysis, isothermality (bio3), max temperature of warmest month (bio5), min temperature of coldest month (bio6), mean temperature of coldest quarter (bio11), precipitation of driest month (bio14) and precipitation of wettest quarter (bio16), and elevation (alt) are the key environmental factors that affect the distribution of C. nigricornis.
| Code | Percent contribution (%) | Permutation importance (%) |
|---|---|---|
| Bio3 | 2.1 | 14.9 |
| Bio5 | 4.2 | 20.4 |
| Bio6 | 18.3 | 39.7 |
| Bio11 | 10.4 | 5.3 |
| Bio14 | 11.2 | 4.1 |
| Bio16 | 44.5 | 10.5 |
| Alt | 9.3 | 5.1 |
The key environmental variables affecting the distribution of C. nigricornis are precipitation and temperature (Figure 3). In terms of precipitation, when precipitation of driest month (bio14) is < 15 mm or more than 74.3 mm, the existence probability of C. nigricornis is < 0.33, and the optimum value is 24.5 mm. According to the classification method of suitable areas, precipitation of driest month (bio14) in the area above the moderately-suitable area is more than 710.28 mm. In terms of temperature, the optimum min temperature of coldest month (bio6) is 12.5°C, which is lower or higher than this value will reduce the existence probability of C. nigricornis. When the value is 6.1°C–15.6°C, the existence probability of C. nigricornis is > 0.33. Similarly, when the existence probability of C. nigricornis is > 0.33, mean temperature of coldest quarter (bio11) is 10.3°C–19.7°C, and the optimum value is 16.6°C.
3.3 Potential Suitable Under Current Scenario
According to the potential geographical distribution pattern (Figure 4) under current scenario, it can be seen that southeastern China and part of southwestern China were mainly suitable areas for C. nigricornis. Compare the prediction result of the two models, the suitable areas all contain Jiangxi, Hunan, Zhejiang and Anhui in the lower-middle reaches of the Yangtze River, southern Yunnan, Taiwan, Hainan, Guangdong and Guangxi.
From the result of MaxEnt model (Figure 4A), it can be seen that the suitable habitat area of C. nigricornis in China is 128.91 × 104 km2, accounting for 13.40% of the total land. The highly-suitable areas are scattered in several provinces in southern Yangtze River, whereas the moderately-suitable areas are divergent distribution with taking as the center of Poyang Lake plain and Lianghu plain in Xiang-gan area, including the southeastern Chongqing, the central of Hunan, the eastern of Jiangxi, the southern of Anhui, the western of Zhejiang, the northern of Fujian, the most regions of Guangdong, the northwest of Guangxi and the southern of Yunnan. The area of the highly-suitable areas, the moderately-suitable areas and the lowly-suitable areas account for 2.05%, 5.29% and 6.08% of the total land, respectively. From the result of GARP model (Figure 4B), it can be seen that the suitable habitat area of C. nigricornis in China is 166.09 × 104 km2, accounting for 17.27% of the total land. The highly-suitable areas are mainly distributed in Guangdong-Guangxi region to the Eo-Wan region, which occupies most of the area of several provinces south of the Yangtze River. The moderately-suitable areas are mainly distributed in the southern of Yunnan and Guizhou. The area of the highly-suitable areas, the moderately-suitable areas and the lowly-suitable areas account for 12.25%, 3.0% and 2.06% of the total land, respectively.
3.4 Potential Suitable Under Future Scenarios
According to the results of the MaxEnt model (Figure 5A–F and Table 3), we can know that the changes of suitable area of C. nigricornis are different from current to the 2050s and from the 2050s to the 2090s under different climate scenarios. The suitable areas of C. nigricornis will decrease by 3.19% in the 2050s and then increase by 12.04% in the 2090s under the SSP1-2.6 scenario. It will increase by 5.79% in the 2050s and then decrease by 7.53% in the 2090s under the SSP2-4.5 scenario, whereas it will increase by 33.03% in 2050s and then decrease by 23.31% in the 2090s under SSP5-8.5 scenario. Although the situation and time are different, but the suitable areas are still concentrated in the Yangtze-Huaihe river basin, the lower-middle reaches of the Yangtze River, South China and Southwest China. Similarly, from the results of the GARP model (Figure 5G–L and Table 3), we can know that the suitable areas of C. nigricornis will increase by 8.41% in 2050s and then continue to increase by 6.11% in 2090s under SSP1-2.6 scenario. It will increase by 23.84% in the 2050s and then decrease by 0.88% in the 2090s under the SSP2-4.5 scenario, whereas it will increase by 34.37% in 2050s and then decrease by 1.75% in 2090s under SSP5-8.5 scenario. As climate scenarios and time change, the boundaries of suitable areas will expand to the north and southwest of China, the most obvious provinces are Sichuan, Chongqing and Yunnan.
| Model type | Scenario, times | Unsuitable area | Lowly-suitable area | Moderately-suitable area | Highly-suitable area |
|---|---|---|---|---|---|
| MaxEnt | Current | 832.76 | 58.39 | 50.76 | 19.76 |
| SSP1-2.6, 2050s | 836.89 | 44.68 | 58.49 | 21.62 | |
| SSP2-4.5, 2050s | 825.31 | 61.90 | 51.63 | 22.84 | |
| SSP5-8.5, 2050s | 790.18 | 44.16 | 111.30 | 16.03 | |
| SSP1-2.6, 2090s | 821.86 | 53.75 | 65.43 | 20.64 | |
| SSP2-4.5, 2090s | 835.58 | 43.09 | 65.95 | 17.06 | |
| SSP5-8.5, 2090s | 830.16 | 59.05 | 53.45 | 19.02 | |
| GARP | Current | 795.51 | 19.73 | 28.79 | 117.57 |
| SSP1-2.6, 2050s | 781.56 | 26.81 | 28.97 | 124.27 | |
| SSP2-4.5, 2050s | 755.91 | 32.45 | 40.43 | 132.80 | |
| SSP5-8.5, 2050s | 738.42 | 35.10 | 54.97 | 133.11 | |
| SSP1-2.6, 2090s | 770.55 | 29.30 | 41.01 | 120.75 | |
| SSP2-4.5, 2090s | 757.72 | 27.90 | 40.17 | 135.81 | |
| SSP5-8.5, 2090s | 742.33 | 43.49 | 26.47 | 149.31 |
For the highly-suitable areas from MaxEnt model, it can be seen that these regions is small and scattered in southern China. Under SSP1-2.6 scenario, its area will increase by 1.86 × 104 km2 from current to the 2050s, whereas it will decrease by 0.98 × 104 km2 from the 2050s to the 2090s. Under SSP2-4.5 scenario, its area will increase by 3.08 × 104 km2 from current to the 2050s, whereas it will decrease by 5.78 × 104 km2 from the 2050s to the 2090s. Under SSP5-8.5 scenario, its area will decrease by 3.73 × 104 km2 from current to the 2050s, whereas it will increase by 2.99 × 104 km2 from the 2050s to the 2090s. It can be seen that the future climate change will promote the expansion of the highly-suitable area of C. nigricornis, but not infinite expansion. At the same time, the area of the highly-suitable areas will have an adaptation process to the sudden warming of climate. Similarly, it can be seen from GARP model that these regions is big and concentrated in southern China. Under SSP1-2.6 scenario, its area will increase by 6.7 × 104 km2 from current to the 2050s, whereas it will decrease by 3.52 × 104 km2 from the 2050s to the 2090s. Under SSP2-4.5 scenario, its area will increase by 15.24 × 104 km2 from current to the 2050s, whereas it will increase by 3.01 × 104 km2 from the 2050s to the 2090s. Under SSP5-8.5 scenario, its area will increase by 15.55 × 104 km2 from current to the 2050s, whereas it will increase by 16.2 × 104 km2 from the 2050s to the 2090s. From the predicted results of this model, the impact of future climate change on C. nigricornis is consistent with the predicted results of Maxent model, that is, the future climate change will promote the expansion of the highly-suitable areas, and the area of the highly-suitable areas will have an adaptation process to the sudden warming of climate.
For the moderately-suitable areas from MaxEnt model, it can be seen that these areas are still distributed in southern China, but the area is twice more than that of the highly-suitable areas. Under SSP1-2.6 scenario, its area will increase by 7.73 × 104 km2 from current to the 2050s, whereas it will increase by 6.95 × 104 km2 from the 2050s to the 2090s. Under SSP2-4.5 scenario, its area will increase by 0.88 × 104 km2 from current to the 2050s, whereas it will increase by 14.31 × 104 km2 from the 2050s to the 2090s. Under SSP5-8.5 scenario, its area will increase by 60.54 × 104 km2 from current to the 2050s, whereas it will decrease by 57.85 × 104 km2 from the 2050s to the 2090s. It can be seen that the moderately-suitable area decreased from 2050s to 2090s under SSP-8.5 scenario, whereas the other scenarios and time periods all increased. It also means the future climate change will be more conducive to the expansion of the moderately-suitable areas. Similarly, it can be seen from GARP model that these regions are mainly distributed in the northern part of the highly-suitable area. Under SSP1-2.6 scenario, its area will increase by 0.18 × 104 km2 from current to the 2050s, whereas it will increase by 12.04 × 104 km2 from the 2050s to the 2090s. Under SSP2-4.5 scenario, its area will increase by 11.64 × 104 km2 from current to the 2050s, whereas it will decrease by 0.26 × 104 km2 from the 2050s to the 2090s. Under SSP5-8.5 scenario, its area will increase by 26.18 × 104 km2 from current to the 2050s, whereas it will decrease by 28.5 × 104 km2 from the 2050s to the 2090s.
4 Discussion
Based on MaxEnt model and GARP model, the paper predicted the suitable areas of C. nigricornis in China, and analyzed the appropriate distribution degree of C. nigricornis in different regions. It was found that the prediction results of two models could reflect the distribution of C. nigricornis well, but the AUC value of MaxEnt model were higher than that of GARP model. The distribution area predicted by GARP model is wider, which can better reflect the range of potential distribution area, but it may have a high false positive rate, while the result predicted by MaxEnt model is conservative, but may be more accurate (Cui et al. 2016; Zhang et al. 2016). Therefore, combining the predictions of MaxEnt and GARP can take advantage of each other's strengths to gain a more complete picture of the potential distribution of C. nigricornis.
Insects are very sensitive to changes in the external environment such as temperature, precipitation or humidity (Cai, Pang, and Yang 2001), so climate change will have an important impact on the survival and development of insects. Temperature variation will affect the growth and development of insects (overwintering survival rate), metabolic rate, survival rate, reproduction rate and other life activities, whereas precipitation changes have direct and indirect effects on insects, such as heavy rainfall directly affects the population size of small insects or indirectly affects insects by changing the relative humidity of the air (Guo, Ma, and Kang 2023; Youngblood et al. 2023). According the study, the important factors affecting the distribution of C. nigricornis are min temperature of coldest month (bio6), mean temperature of coldest quarter (bio11), precipitation of driest month (bio14) and precipitation of wettest quarter (bio16). C. nigricornis overwinter with eggs, and begin to hatch in late April. This period of time coincides with the coldest season, so min temperature of coldest month and mean temperature of coldest quarter determines whether it can survive the winter and hatch safely. C. nigricornis is an insect that draw water from plants and do not need additional water. So as long as there is vegetation, drought will have no effect on its growth. In contrast, dry conditions are more conducive to incubation (Wu, Hao, and Kang 2021). In southern China, the dry season is usually from February to April, when locust eggs hatch. Therefore, precipitation of wettest quarter (bio14) is important factors affecting the survival of C. nigricornis (Chang et al. 2008). Similarly, June to August is the key period for adult emergence, and insufficient precipitation will increase the risk of water loss of larva (Dang and Chen 2011).
Ceracris nigricornis is a host-oriented leaf-eating pest, its growth, development and distribution are closely related to host species (Zhang et al. 2002). Hunan, Jiangxi, Fujian, Zhejiang, Anhui, Sichuan, Zhejiang and Guangdong are the main production areas of bamboo resources (Li and Feng 2019), this is consistent with the predicted distribution of C. nigricornis in this study. From the prediction results of MaxEnt model and GARP model under current scenario, the highly-suitable areas for C. nigricornis are the southern part of Anhui, the western part of Zhejiang, the northeastern part of Guangxi, the eastern of Hainan and Taiwan. Therefore, it is suggested that the local forestry authorities should strengthen the monitoring of C. nigricornis to prevent it from causing serious harm to bamboo forests. In future climate scenarios, it can be seen that climate warming will promote the expansion of the suitable area, but not infinite expansion. The potential distribution of C. nigricornis in China is still in the south of Qinling-Huaihe River, which may indicate that the low temperature in winter in the northern region is still in the range that restricts the migration of C. nigricornis to the north. To the new highly-suitable or moderately-suitable areas of province such as Sichuan, Chongqing and Yunan, should be strengthened to quarantine inspection and pest monitoring in combination with the spatial and temporal distribution of bamboo.
In this study, MaxEnt model and GARP model were used to predict the suitable habitats of C. nigricornis in China by combining bioclimatic variables, elevation and actual site data. However, there are still some problems. For example, the impact of soil factors on food sources for food moths has not been fully considered, and the area of non-growing areas such as water systems and towns has not been eliminated. This may inevitably increase the niche width. Meanwhile, some parameters in this study are default, which may lead to low expressive ability of the model (Sillero et al. 2021). Therefore, we will strengthen this consideration to make the prediction of the potential distribution of species more realistic in future studies.
Author Contributions
Chun Fu: data curation (equal), formal analysis (equal), investigation (equal), software (equal), validation (equal), visualization (equal), writing – original draft (lead). Xuanye Wen: formal analysis (equal), investigation (equal), resources (equal), validation (equal), writing – review and editing (equal). Tingting Huang: data curation (lead), methodology (equal), software (equal), validation (equal). Yanli Wang: data curation (equal), software (equal), validation (equal). Xiu Liu: data curation (equal), investigation (equal), software (equal). Na Jiang: formal analysis (equal), validation (equal). Rulin Wang: conceptualization (equal), project administration (equal), supervision (equal), visualization (equal), writing – review and editing (equal). Jinpeng Zhao: conceptualization (equal), methodology (equal), resources (equal), supervision (equal), visualization (equal), writing – review and editing (equal).
Acknowledgements
This study was funded by opening Foundation of Key Laboratory of Sichuan Province for Bamboo Pests Control and Resource Development (ZLKF202305, ZLKF202202), Science and Technology Program of Leshan Normal University (2022SSDJ005, KYPY2023-0006, XJR17005 and LZD010).
Conflicts of Interest
The authors declare no conflicts of interest.
Open Research
Data Availability Statement
All data are in the main text or supporting information.
