Monitoring and Management of agricultural insect pest with GIS

Compiled for GEOG 560, Fall 2022, Oregon State University by Grace Tiwari

Author’s Notes: As an agricultural science student working primarily to ensure food security and sustainable agriculture development, I implement strategiesaimed at insect pest management. UNFAO estimates that annually between 20 to 40 percent of global crop production are lost to pests. Geographic InformationSystems (GIS) is integral to make decisions for the insect pest control through timely planning and correct information. The following annotated bibliographydocuments a variety of GIS based tools and techniques that have been used to map and monitor crop pest infestation in order to help researchers and farmers totake proper management planning and decisions. Please submit any questions or comments through the contact form at the bottom.

Annotated Bibliography

Carriere, Y., Ellsworth, P.C., Ellers-Kirk, C., Barkley, V., Antilla, L., 2006. A GIS-based approach for areawide pest management: The scales of LygusHesperus movement to cotton from alfalafa, weeds, and cotton. Entomologia Experimentallis et Applicata, 118(3):203 – 210. DOI:10.1111/j.1570-7458.2006.00384.x

It is important to understand the effect of cropping patterns on population dynamics, dispersal, and habitat selection of insect pests. In this study, the authorsincorporated a general approach based on global positioning system (GPS) and geographic information system (GIS) technologies combined with spatial statisticsto assess and map the maximum distance at which three different plant species (alfalfa, weeds and cotton) affects lygus bug population density. The model showedthat Lygus hesperus population density was significantly lower in cotton than in the other field types, confirming that cotton is not a suitable or preferred host forplant bugs in early summer. GIS was used to draw four concentric 750-m-wide rings around each pheromone trap to measure area of all Bt and non-Bt cottonfields within each ring. Hence, these outcomes from GIS based approach can substantially contribute to the development of areawide pest management.

Dawn M Olson, Kristina K Prescott, Adam R Zeilinger, Suqin Hou, Alisa W Coffin, Coby M Smith, John R Ruberson, David A Andow, LandscapeEffects on Reproduction of Euschistus servus (Hemiptera: Pentatomidae), a Mobile, Polyphagous, Multivoltine Arthropod Herbivore, EnvironmentalEntomology, Volume 47, Issue 3, June 2018, Pages 660–668, https://doi.org/10.1093/ee/nvy045

This paper describes a relatively simplistic GIS approach to manage brown stick bug (an economic pest in several crops) by understanding the relative influence oflocal and landscape factors on their reproduction. GIS was used to determine the percentage area in the landscape that contained the insect host crops (maize,peanut, cotton and soybean). All geospatial manipulations and analyses were conducted using ArcGIS. The authors found soybean as most important variable fordetermining brown stink bug population increase with much greater reproduction in soybean fields than in other crop species. The article demonstrates how pestreproduction can be monitored and compared among different landscapes using GIS tool.

Ji Rong, Li Dian-Mo, Xie Bao-Yu, Li Zhe, Meng Dong-Li, Spatial Distribution of Oriental Migratory Locust (Orthoptera: Acrididae) Egg PodPopulations: Implications for Site-specific Pest Management, Environmental Entomology, Volume 35, Issue 5, 1 October 2006, Pages 1244–1248,https://doi.org/10.1093/ee/35.5.1244

In this article, Rong et al describes the implications for site-specific pest management strategies by analyzing the spatial distribution of oriental migratory locustegg pod population. Each sample location was georeferenced with a hand-held global positional system (GPS) unit (Magellan GPS315; Magellan Systems, SanDimas, CA and Garmin map 60cs; Garmin International, Lenexa, KS) in the Universal Transverse Mercator (UTM) coordinate system to be used in geostatisticalanalysis. ARCGIS version 8.1 was used to create coverages of oviposition sites, kriged values, and elevation. The results of this study showed two recurringpatterns of the spatial distribution of locust egg pods: patchiness and aggregation. These models could provide critical information needed for the implementationof site-specific management of egg pod populations such as spring plowing and autumn harrowing. Moreover, this approach can be implemented to provide a basefor the development of targeted monitoring and spraying programs during peak hatch, which would reduce the pesticide load and therefore minimize the negativeimpact on nontarget species and the environment.

Vanegas, F.; Bratanov, D.; Powell, K.; Weiss, J.; Gonzalez, F. A Novel Methodology for Improving Plant Pest Surveillance in Vineyards and Crops UsingUAV-Based Hyperspectral and Spatial Data. Sensors 2018, 18, 260. https://doi.org/10.3390/s18010260

This paper describes a remote sensing-based methodology for developing a predictive detection method for pests. The authors implemented the methodology toincrease the efficiency of existing surveillance practices (human inspectors and insect traps) for the development of a predictive model for phylloxera detection ingrape vineyards. The methodology uses UAV (Unmanned Aerial Vehicles) integrated with processed airborne RGB, multispectral, hyperspectral, and ground-collected data. The methods used for processing the data are the photogrammetry of RGB imagery; the development of an airborne RGB-based digital elevationmodel in order to assess the canopy vigour; the georeferencing of different datasets; the processing of multispectral and hyperspectral imagery to generatevegetation indices, and the extraction of mean spectral signatures of grapevines for different levels of pest infestation at the canopy level. The results show thathyperspectral imagery has the potential to detect grape phylloxera before it is apparent to visual inspection. The authors presented the results of comparing adigital vigour model of the vineyard to an expert visual assessment. They found that the two assessments correlate positively indicating that the developed methodis a correct approach for generating vigour assessments in vineyards. The authors also identified mean spectral signatures for different levels of infestation for theChardonnay variety at two different times of the year which helped them find regions of interest in the spectrum in order to generate new vegetation indices tohighlight grape phylloxera infestation. The methods, workflow, results and analysis presented in this research can contribute to the generation of valuableinformation for plant pest surveillance. The presented methodology could also be extrapolated to other areas of research in remote sensing, such as mineralsexploration, biodiversity, and ecological assessment.

Franke, J., Menz, G. Multi-temporal wheat disease detection by multi-spectral remote sensing.Precision Agric 8, 161–172 (2007).https://doi.org/10.1007/s11119-007-9036-y

In this article, the authors examine the potential of multi-spectral remote sensing for a multi-temporal monitoring of fungal wheat infection. Three high-resolutionremote sensing images were used to execute a spatio-temporal analysis of the infection dynamics. A decision tree, using mixture tuned matched filtering (MTMF)results and the Normalized Difference Vegetation Index (NDVI), was applied to classify the data into areas showing different levels of disease severity. The resultsshowed that high-resolution multi-spectral data are generally suitable to detect in-field heterogeneities of crop vigor but are only moderately suitable for earlydetection of crop infections due to the high misclassification rate observed in this study at early growth stages. This indicates the suitability of these data for thedetection of diseases with late occurrence and/or high infection growth rates. However, the low temporal resolution of current sensor systems with high spatialresolution is a restrictive factor for practical implementation. Remote sensing images can be valuable as a source of additional information for the generation offungicide application maps, particularly at growth stages when fungal infections are to be expected.

Hu, C.; Kong, S.; Wang, R.; Zhang, F. Radar Measurements of Morphological Parameters and Species Identification Analysis of Migratory Insects.Remote Sens. 2019, 11, 1977. https://doi.org/10.3390/rs11171977

Migratory insect identification has been concerning entomologists for a long time. In this paper, the author reports an improved measurement method of insectmass and body length using an experimental multi-frequency radar. Previously, the nocturnal behavior and very small radar cross-sections (RCS) of migratoryinsects has made individual detection challenging. In this paper, the authors report an improved measurement method of insect mass and body length using a highfrequency radar. They evaluated the classification performance using independent features and found that insect mass is the most effective feature fordiscriminating different insect species. These findings provided improvements in individual parameter measurement for insect radars and imply the possibility ofhigher precision species identification, which would promote the biomass quantification of over-flying migratory insects to enable studying the role of insectmigration in ecosystems and pest control.

Marković D, Vujičić D, Tanasković S, Đorđević B, Ranđić S, Stamenković Z. Prediction of Pest Insect Appearance Using Sensors and Machine Learning.Sensors (Basel). 2021 Jul 16;21(14):4846. doi: 10.3390/s21144846. PMID: 34300586; PMCID: PMC8309862.

In this article, the authors presented a machine learning based prediction model that indicates the possibility of pest appearance using the temperature and relativehumidity as environmental parameters. Several machine learning algorithms for classification were applied and their accuracy for the prediction of insectoccurrence was presented (up to 76.5%). In the case of a period of five days, the ML model with AdaBoost algorithm stood out with highest accuracy of theaffected detections, and the percentage of false detections was 11%. The proposed ML model can indicate the potential situation in the field and provide farmersan optimal platform for work. In addition, the farmers can better plan their activities ahead of time (a few days or weeks). With this model, farmers can assesstheir priorities, make informed decision on which sites to visit first, as well as to postpone the application of insecticides in case precipitation is expected.

Willers, J.L., Jenkins, J.N., Ladner, W.L. et al. Site-specific Approaches to Cotton Insect Control. Sampling and Remote Sensing Analysis Techniques.Precision Agric 6, 431–452 (2005). https://doi.org/10.1007/s11119-005-3680-x

In large commercial cotton fields, it is difficult for observers on the ground to map habitat boundaries and estimate the abundance of varied insect populationdensity. It is important to consider the influence of habitat structure while sampling multiple population densities, to correctly estimate the mean abundance forbetter pest management decisions. The authors of this article suggest remote sensing and GIS can efficiently create geo-referenced, stratified maps of cotton fieldhabitats. The imagery used in this project was obtained from an airborne, digital camera system [ITD-Spectral Visions RDACS (Remote Data Acquisition CameraSystem) Stennis Space Center, MS, USA] that acquired three bands (540±5, 695±5 and 840±5 nm) at 2 m spatial resolution. The imagery was geo-referenced toGPS ground control points, using nearest neighbor resampling, to the Universal Transverse Mercator (UTM) coordinate system. From the multi-band composite,the normalized difference vegetation index (NDVI) was determined for each pixel in the image frame using ERDAS ImagineÒ software (Leica Geosytems(ERDAS), Atlanta, GA, USA).

Kirkeby C, Rydhmer K, Cook SM, Strand A, Torrance MT, Swain JL, Prangsma J, Johnen A, Jensen M, Brydegaard M, Græsbøll K. Advances inautomatic identification of flying insects using optical sensors and machine learning. Sci Rep. 2021 Jan 15;11(1):1555. doi: 10.1038/s41598-021-81005-0.PMID: 33452353; PMCID: PMC7810676.

It is important for farmers to understand the location and time of pest infestation so that while spraying pesticides, they are targeted to pests only and not thebeneficial insects. The authors in this article suggest that a promising solution to this issue could be optical sensors combined with machine learning. In this study,authors targeted the analyses on nine species and used optical sensor that transmitted Ø2″ collimated linearly polarized laser beam of both 808 nm and 980 nm andflight activity of insects were recorded. The results showed that optical signals can be processed using machine learning to distinguish between taxonomic groupsof flying insects. Optical sensor and machine learning can add value to precision farming as it can help in early detection of pests as well as rapid identification oflocal hotspots of pests. The authors also expect that combined with other features such as modulation spectra and fluorescence, remote insect recognition willenable to optimize the application of insecticides in space and time.

Tabuchi, K., Murakami, T., Okudera, S., Furihata, S., Sakakibara, M., Takahashi, A., Yasuda, T., Predicting potential rice damage by insect pests usingland use data: A 3-year study for area-wide pest management. Agriculture, Ecosystems & Environment. 2017; 249(4-11). Doi: 10.1016/j.agee.2017.08.009

In this article, the authors tested whether the occurrence of pecky rice damage caused by sorghum plant bug could be predicted using a spatial model based on landuse data. ArcGIS software was used to create maps for visualizing potential priority areas for the allocation of available pest management resources. Linear mixedmodel analysis and model selection by Akaike’s information criterion was used to evaluate pecky rice damage. The model was not sensitive enough forquantitative prediction but was accurate to predict whether brown rice would be first grade or not. Further validation of the spatial predictive model is needed inother regions to determine model’s limiting factors and constraints.

Dong, Y., Xu, F., Liu, L., Du, X., Ren, B., Guo, A., Geng, Y., Ruan, C., Ye, Hu., Huang, W., Zhu, Y., Automatic System for Crop Pest and Disease DynamicMonitoring and Early Forecasting. IEEE Journal of selected topics in applied earth observations and remote sensing. 2020; Vol 13(4410-4418). DOI10.1109/JSTARS.2020.3013340

In this article, the authors selected the sensitive indexes for pest and disease habitat monitoring and early forecasting, and then optimized the forecasting model’sparameters to develop an automatic system based on web GIS platform. The authors used spatial and attribute data to monitor and forecast wheat yellow rustdisease and locust habitat. For spatial data, they include land-use data with 30 m spatial resolution, administrative division data, wheat planting area, locust area,and remote sensing data including MODIS surface reflectivity (MOD09A1), normalized difference vegetation index (NDVI) (MOD13A2), land surfacetemperature (LST) (MOD11A2), landsat and sentinel images. Automatic system approach was found to be efficient in pest and disease management andprevention by improving accuracy of monitoring and forecasting. We can also recommend this system for recommending appropriate application of chemicalpesticides for farmer’s profitability and environment sustainability.

Zhou, & Li, X. (2021). The monitoring of chemical pesticides pollution on ecological environment by GIS. Environmental Technology & Innovation., 23.https://doi.org/10.1016/j.eti.2021.101506

Chemical pesticides are being extensively used on agricultural produce to increase productivity and profitability. However, it is important to comprehend andmonitor the situation and promote strict supervision and management of pesticide use to protect human health, biodiversity and environment. In this article,authors recommend the use of GIS technology to connect with the Internet and establish a monitoring system for the pollution of chemical pesticides to theecological environment. Authors used a system that analyzed the use of pesticides, as well as the areas and trends of use, and presented comprehensive and reliableresults and information. This process requires the principles and methods of system design to be strictly followed, and the use of remote sensing image technologyto establish a model to obtain the basic situation of agricultural landscape for effective pesticide application decision making.

Rafoss, T. Spatial stochastic simulation offers potential as a quantitative method for pest risk analysis. Risk Analysis. 2003; Vol 23 (651-661). https://doi-org.oregonstate.idm.oclc.org/10.1111/1539-6924.00344

Pest risk analysis is used to evaluate and assess the potential risks of the introduction and establishment of plant pests into a new geographic location. In thisarticle, author combines current quantitative methodologies, stochastic simulation, and GIS with knowledge of pest biology and environmental data to derive newinformation about pest establishment potential in a geographical region where a pest had not been introduced. Stochastic simulation (e.g., in a Monte Carlosimulation of a complete pest risk model) was used to sample the distribution defined by the GIS model. Regardless of its potential, GIS has so far had relativelylimited use in risk analysis, which could be due to the absence of facilities in GIS for handling uncertainty in spatial data. Further research on data errormanagement and error visualization is needed for complete pest risk analysis efficacy.

Setiyono, T.D., Quicho, E.D., Gatti, L., Campos-Taberner, M., Busetto, L., Collivignarelli, F., Garcia-Haro, FJ., Boschetti, M., Khan, NI., Holecz, F.Spatial Rice Yield Estimation Based on MODIS and Sentinel-1 SAR Data and ORYZA Crop Growth Model. Remote Sensing. 2018; Vol 10.2 (293). DOI:10.3390/rs10020293

Smallholder farmers are vulnerable to the risks from pest and disease outbreak, extreme weather events and market shocks that threaten their economy and foodsecurity. Crop insurance can be a viable solution to address this issue. This paper demonstrates and validates the remote sensing methodology and estimate cropgrowth model-based rice yield for its application in crop insurance. The authors developed a system that combined MODIS and SAR-based remote-sensing data togenerate spatially explicit inputs for rice using a crop growth model. The authors demonstrated a synergetic method of exploiting optical and SAR remote-sensingdata in the context of assimilation of critical information such as rice area, start of season, and LAI for rice yield simulation. This approach shows itself to be wellsuited for reconstructing historical yield data in the context of crop insurance programs. The information from the use of this technology can help to support riskreduction efforts by supporting crop insurance, identify and target hot-spot areas for interventions to reduce yield gaps optimize resource use to ensure sustainablerice production.

Deleon, L., Brewer, M.J., Esquivel, I.L., Halcomb, J. Use of a geographic information system to produce pest monitoring maps for south Texas cotton andsorghum land managers. Crop Protection. 2017; 101(50-57). http://dx.doi.org/10.1016/j.cropro.2017.07.016

Understanding the economic threshold to monitor pest infestation in a field is an important pest management practice. In this article, GIS software ArcMap wasused to create pest monitoring map that displayed variation of economic thresholds (below or above), visualized by color-coding of insect densities acrosssampling sites. GIS tools were used to collect both insect abundance and plant injury measurements. Before entering the field, authors created a basemap wasusing an orthorectified image that included demonstration area, taken by the National Agriculture Imagery Program (Geospatial Data Gateway). Then a GPSenabled handheld device Trimble Juno 3B was used to run GIS ArcPad. Overall, this approach can be recommended for insecticide use decision making.

Monitoring and Management of agricultural insect pest with GIS

Annotated Bibliography

Additional Resources:

Comments

Leave a Reply Cancel reply