Prioritization of absent quarantine pests in Brazil through the Analytical Hierarchy Process

Abstract Introducing an Absent Quarantine Pest (AQP) can cause severe economic, social and environmental impacts, generating food insecurity. The Analytical Hierarchy Process (AHP) method is an excellent tool for prioritizing APQs, allowing countries to better prepare against these threats. This study aimed to determine which AQPs should be prioritized in Brazil. For this, 20 AQPs were selected from the Brazilian official list. The selection was based on pests intercepted by Brazil between 2015 and 2018 and by countries of the European and Mediterranean Plant Protection Organization, in the international movement of plants. It can be concluded that out of the 20 AQPs studied, 17 are the priority and that the AHP method is effective for this purpose. Other countries from different continents can use this methodology to prioritize PQAs and thus create strategic plans to prevent entry into their territories and economic, social, and environmental impacts.


Introduction
Introducing an Absent Quarantine Pest (AQP) can cause severe economic, social and environmental impacts, generating food insecurity.In some cases, it reaches the level of a trade barrier for the country involved (WTO, 2020).Worldwide costs with invasions reached US$ 1.288 trillion from 1970 to 2017, and the average annual cost was US$ 26.8 billion.In Brazil, the minimum estimated costs for just 16 invasive species were US$ 105.3 billion over the last 35 years , with an average cost of US$ 3.02 (± 9.8) billion per year (Adelino et al., 2021).Brazilian agriculture is the sector most impacted by biological invasions, with an estimated economic cost of US$ 39.61 billion (Adelino et al., 2021).This study is important to reduce production losses to avoid pressure on conserved areas and, consequently, combine sustainable development with environmental protection.
Brazil is responsible for producing food that serves 800 million people from different countries.By 2050, national grain production could exceed 500 million tons, being even more critical for world food security (Canal Rural, 2020).Despite the volumes produced and sold, some obstacles hamper production, especially pests.Among the primary means of spreading pests is the increase in the flow of people around the world (IPPC, 1997), which reached U$ 1.5 billion arrivals (ONU, 2022).Furthermore, the increase in commercial relations contributes to potentially introducing these harmful organisms (IPPC, 1997).

Selection of AQPs
In this study, we used the AHP method adopted by Laranjeira et al. (2018) to assess 20 AQPs not included in the official list of priorities listed in Ministério da Agricultura e Pecuária (MAPA) Ordinance 131 (Brasil, 2019).
Among the hundreds of AQPs that make up the official Brazilian list (Brasil, 2018), twenty AQPs were selected according to the following parameters: I -Pests with at least five interceptions recorded between 2015 and 2018 by MAPA and the European and Mediterranean Plant Protection Organization -EPPO.The records of interceptions were made available by the Department of Plant Health -DSV/MAPA, II -Pests with at least five host crops and at least one of these crops with an area implanted in three Brazilian regions, according to the Instituto Brasileiro de Geografia e Estatística (IBGE), III -Pests occurring in at least three countries with commercial relations with Brazil (Table 1).

Criterion and sub-criteria
The AHP is composed of the entry, establishment, spread, d and, estimated impacts criteria.These criteria have 19 correlated sub-criteria, each with its importance.This importance is characterized by the adoption of different weights between the sub-criteria (Table 2).
It has six sub-criteria referring to the probability of entry of an AQP into the Brazilian territory.Mainly related to the geographic distribution of AQPs and the commercial movements of their hosts (Adaime et al., 2018).
Refers to the potential for establishment and spread of absent quarantine pests after occasional entry into the Brazilian territory.Its seven sub-criteria are directly related to climate, area of cultivation of host plant species, human intervention, and biological characteristics (Fidelis et al., 2018).
The Estimated impacts criterion has six sub-criteria related to the possible economic, social, and environmental impacts caused by an eventual introduction of AQP into the Brazilian territory.These impacts can be directly due to the product's unfeasibility and indirect effects caused by the social and environmental issues involved (Lohmann et al., 2018).
After selection, the 20 AQPs were evaluated by the AHP method and subjected to each of the 19 sub-criteria that are part of the entry, establishment, and spread and estimated impacts criteria, reaching scores that ranged from 0 to 1,000 (Hilman et al., 2023).

Statistical analysis
Data from the 20 AQPs evaluated by the AHP were subjected to basic statistical analysis and the K-Means Clustering algorithm.This method, improved by Hartigan (1975), classifies data sets with similar results, regardless of presets, forming groups or clusters.This method classifies data sets with similar results, independent of presets, forming groups or clusters.The organization analyzes possible clusters using a multivariate data mining technique, automatically homogenizing the data into groups according to their similarities (PUC-RIO, 2020).(Hilman and Goulart, 2015).For example, H. armigera has the potential to cause about 40% production losses.Soybean, corn, and cotton represent some crops of choice and are among the main species grown in Brazil (CEPEA, 2019).The most recent pests introduced in the country were Drosophila suzukii Matsumura (Diptera: Drosophilidae) in 2014, which spread to Uruguay, Argentina, and Chile (Garcia, 2020;Garcia et al., 2022), and Sternochetus mangiferae (Fabricius) (Coleoptera: Curculionidae) in 2017 (Silva and Ricalde, 2017).
I view the damage caused by AQP to food-producing countries, and preventive measures are taken to reduce the risk of introduced pests.Brazil has prevention strategies established in international agreements.Among them are pest risk analysis, phytosanitary certification, global agricultural surveillance, plant quarantine, and specialized laboratory support (IPPC, 1997).The Ministry of Agriculture, Livestock, and Supply -MAPA established as AQPs in about 600 organizations widely distributed worldwide (Brasil, 2018).These pests have biological characteristics and different forms of dissemination that can find conditions for their establishment in the national territory.Brazil has a vast area dedicated to agriculture and marked environmental diversity (IBGE, 2019), which can facilitate this dissemination.
In this context of threats and impacts, it is necessary to establish methodologies on pest prioritization for risk and human and financial resources management.The Analytical Hierarchy Process (AHP) method makes it possible to remove subjectivity from decisions, logically structuring a complex problem and facilitating the determination of priorities (Saaty, 2013).
The Analytical Hierarchy Process (AHP) can be used for several purposes, such as: to evaluate alternative pest-control strategies (Wan et al., 2009), for application in the selection of plant types on the community agroforestry land (Rahmawaty et al., 2022), for selection of agricultural irrigation systems (Veisi et al., 2022), for prioritization of barriers to offshore wind energy (Dhingra et al., 2022), for ecological risk assessment of marine microplastics (Zhang et al., 2022), for hospital site selection (Sahin et al., 2019), selecting strategies for rice stem borer management (Abdollahzadeh et al., 2016), among many other applications.
There are few records in the literature focused on pest prioritization using AHP.Therefore, the present study evaluated which AQP should be prioritized by Brazil using the Analytic Hierarchy Process (AHP) methodology.This study evaluates 20 AQPs considered important for MAPA (Brasil, 2019).

Entry criterion
The awned canary grass (Phalaris paradoxa), Common groundsel (Senecio vulgaris), Onionweed (Asphodelus tenuifolius), Chilli thrips (Scirtothrips dorsalis), and Potato cyst nematode (Globodera pallida) presented high risk in the entry criterion (scores more significant than 750).Their prioritization was influenced by the presence in border countries, wide geographic distribution, and considerable import of host material.On the other hand, the corn cyst nematode (Heterodera zeae), rhubarb crown rot (Erwinia rhapontici), omnivorous leaf roller (Platynota stultana), and Pacific spider mite (Tetranychus pacificus) had the lowest risk of entry (scores less than 500).These four AQPs have in common that they are absent in South America and have a restricted worldwide distribution.The other pests reached intermediate scores (between 500 and 750), influenced by geographic distribution and their proximity to Brazil (Table 3).

Establishment and spread criterion
The AQP Spodoptera littoralis (Boisduval), African cotton leafworm, Senecio vulgaris L. groundsel, and Tetranychus pacificus, Pacific spider mite, achieved scores greater than 750, that is, very high potential for establishment and spread.
The pests Acarus siro¸ Flour mite, and Latheticus oryzae, longheaded flour beetle, achieved the lowest scores in this criterion, as they are stored grain pests and do not directly attack crops.All AQPs studied had a high capacity for climatic adaptation, however, the sub-criteria efficiency of control methods and annual natural spread distance estimation had wide variation in scores (Table 4).

Estimated impacts criterion
Most AQPs achieved high or very high scores.The exceptions were Acarus siro and Latheticus oryzae, due to their characteristics of causing low impacts.Scirtothrips dorsalis and Spodoptera littoralis stood out, which reached the maximum scores in all sub-criteria (Table 5).

General result of the prioritization
The prioritization of pests according to the AHP methodology signals the importance of the AQP selected in the present study, especially Scirtothrips dorsalis, Spodoptera littoralis, Senecio vulgaris, Globodera pallida, and Platynota stultana.However, all the pests studied reached expressive results (above 600).The exceptions are the stored grain pests, Latheticus oryzae, and Acarus siro, due to their characteristics already reported in the present study (Table 6).The clustering result (Table 7).
Table 3. Ranking of absent quarantine pests (AQP), their prioritization (P), and score (S), according to the sub-criteria of the entry criterion and sub-criteria: 1, th distance between the nearest location and the border, 2, number of bordering countries in which it occurs, 3, number of countries in which it occurs, 4, number of continents where the pest occurs, 5, the Import volume of the host material, 6, host material import numbers.The final data from the prioritization of pests were subjected to statistical analysis (mean, standard deviation, standard error, and variance) as a prerequisite for applying the K-means algorithm, as there was a similarity in scores of most AQPs.Statistical clustering was necessary to remove subjectivity from data analysis.

AQP
The prioritization of pests according to the AHP methodology [adapted by Laranjeira et al. (2018)] signals the importance of the AQP selected in the present study, especially Scirtothrips dorsalis, Spodoptera littoralis, Senecio vulgaris, Globodera pallida, and Platynota stultana.However, all the pests studied reached expressive results (above 600).The exceptions are the stored grain pests, Latheticus oryza, and Acarus siro, due to their characteristics already reported in the present study.
The final data from the prioritization of pests were subjected to statistical analysis (mean, standard deviation, standard error, and variance) as a prerequisite for applying the K-means algorithm, as there was a similarity in scores of most AQPs.Statistical clustering was necessary to remove subjectivity from data analysis (Table 7).
The K-means algorithm included two, ten, and eight pests in groups 1, 2, and 3, respectively.Group 1 scored close to 600 on the entry criterion, 350 on the establishment and spread criterion, and 480 on the estimated impacts criterion.Group 2 scored approximately 500 on the entry criterion, 700 on the establishment and spread criterion, and 900 on the estimated impacts criterion.Group 3 achieved a score close to 700 in the entry criterion, 700 in the establishment and spread criterion, and 800 in the estimated impacts criterion (Figure 1).

Arabis Mosaic Virus Virus G2
Impatiens Necrotic Spot Virus Virus G2

Discussion
The AQP list prioritized has scored above 500 (Brasil, 2019).In the present study, 17 out of the 20 AQP evaluated reached scores between 687 and 883, affirming that these pests should be officially incorporated into this list.This statement is based on the K-means algorithm for forming groups, from which it was possible to observe that groups 2 and 3 have similar, high, and very high scores, distinguishing in the entry criterion.Pests in group 1 had significantly lower scores and were not considered a priority.The greatest uniformity was found in group 3, formed by Scirtothrips dorsalis, Senecio vulgaris, Globodera pallida, Botrytis fabae, Chondrostereum purpureum, Otiorhynchus sulcatus, Phalaris paradoxa, and Asphodelus tenuifolius.
Scirtothrips dorsalis reached the highest score among the 20 AQPs.Its performance was significant across all criteria, with a maximum score in estimated impacts.In addition to direct impacts, the pest is a virus vector for mango trees and other important host plants (Paula and Oliveira, 2001).The occurrence this pest was found in Anacardium occidentale in a greenhouse in the Ceará state.For this reason, this crop was eliminated (Dias-Pini et al., 2018).Another evidence of the importance of this AQP is its high number of hosts, especially citrus.The impacts of this pest can range from 61 to 74% of host production (Kumar et al., 2013).Brazil is the world's largest producer, with approximately 18 million tons, generating about U$ 2.2 billion (IBGE, 2020).These facts confirm the effectiveness of the method that prioritized S. dorsalis with the highest score because it causes high damage and has a high risk of introduction into the Brazilian territory.
Senecio vulgaris had the third-highest score among the 20 AQPs.It achieved expressive results in the three criteria, evidencing its high capacity for entry, establishment, and spread and impact.Brazilian Normative Instruction in 2020 excludes this species from the official Brazilian AQP list due to the recognition of its presence in Brazil (Brasil, 2020).This situation allows affirming that the method is effective since it determined a high probability of entry of this pest.
Globodera pallida had the fourth-highest prioritization score.This nematode, together with G. rondachiensis, initially evaluated by experts (Laranjeira et al., 2018), can impact the potato and tomato production chains, which have 171,220 hectares of cultivated and about U$ 2.2 billion (IBGE, 2020).These pests can cause approximately 26% damage with an impact of U$ 2.88 billion (Van Oijen et al., 1995).It has records of occurrence in Argentina, Chile, Ecuador, Colombia, and Bolivia (CABI, 2020).MAPA conducts inspections of buses, passenger vehicles, and pedestrians from Bolivia and Peru on the border between Brazil and Bolivia.In these actions, potato tubers (chuños) are frequently intercepted and transported mainly to São Paulo, with a high risk to national agriculture (MAPA, unpublished data).
Botrytis fabae is registered in Colombia, Argentina, and Uruguay, countries bordering Brazil (CABI, 2020).Its wind spread capacity can reach 26 meters per day (Fitt et al., 1985).The importance of this pest suggests the need to design survey and surveillance programs in border areas that have reported its presence.The damage potential in their hosts can reach 50% production (Emeran et al., 2006).Its impacts are also recorded in beans, one of its main hosts, with 2,769,934 hectares cultivated in Brazil.The production value is U$ 1.49 billion (IBGE, 2020).Although the method assigned this pest the 14th highest average in the three criteria, its importance is justified by the high and homogeneous scores.
Chondrostereum purpureum also has an important wind spread capacity, reaching up to 400 km per year (France and Grinberg, 2014).It is present in Argentina, Chile, Paraguay, and Uruguay, with more than 150 hosts, particularly peach, apple, and grape (CABI, 2020).Apples are among its preferred hosts, with a cultivated area of 32,433 ha and U$ 362 million in production value (IBGE, 2020).Its impact can compromise 50% of production (France and Grinberg, 2014).These losses could amount to U$ 181 million.
The AHP prioritized Otiorhynchus sulcatus F. (vine weevil) in the 17th, mainly justified by its high scores with an emphasis on the estimated impacts criterion.This pest has dozens of hosts, such as grapes, strawberries, hops, and Taraxacum officinale Wiggers, and mainly ornamental species, such as violets, roses, primrose, camellia, and begonia (CABI, 2020).This AQP can cause about a 60% reduction in the biomass of the host species (Clark et al., 2012).Larvae attack the lower portion of the stem and roots, making it difficult to control them.Adults feed on leaves and flowers (Fernandez, 2020).It is estimated that the violet production chain encompasses approximately 5,000 producers in Brazil, and losses are estimated at U$ 150 million (Fernandez, 2020).
Concerning Phalaris paradoxa (awned canary grass), the AHP assigned high scores, especially in the entry criterion.This was contextualized with the interception of this pest during an inspection by the International Agricultural Surveillance Service of Foz do Iguaçu, Paraná state (SVA FOZ-VIGIAGRO -MAPA) in a shipment of birdseed from Argentina and confirmed by a laboratory report issued on 16/11/2021, by an accredited laboratory (MAPA, unpublished data).The risk of the probability of entry is based, among other aspects, on the fact that it is present on all continents, especially bordering countries, such as Argentina, Bolivia, Chile, and Uruguay (CABI, 2020).The borders with these countries are recognized as having a high phytosanitary risk because they have numerous crossing points (Spadotto et al., 2014).This species has high seed production, with characteristics of dormancy and emergence periodicity (Taylor et al., 1999).The main impact of this pest occurs in the cultivation of wheat, whose production value in Brazil is U$ 1.06 billion (IBGE, 2020).Losses from this AQP can reach U$ 698 million, equivalent to 86% of production (Taye and Tanner, 1997).
Asphodelus tenuifolius (onionweed) reached the sixth-highest score according to the AHP methodology.It stood out in the three criteria since it is widely distributed across the five continents (CABI, 2020), emphasizing its registration in Bolivia, which has 3,400 km of border with Brazil (IBGE, 2020).It stands out for producing up to 2,300 seeds per plant (Baber et al., 2009).It affects sugar cane, corn, wheat, cotton, and tobacco (Holm et al., 1997).Chickpea production can decrease by 80% and mustard by 56% due to infestation by this pest (Tewari et al., 2001).Losses in sugarcane can be significant due to the 10,109,413 hectares cultivated in Brazil.The production value is U$ 4.6 billion, and losses are estimated at U$ 4.5 billion (IBGE, 2020).
Group 2 is formed by Tetranychus pacificus, Erwinia rhapontici, Rhodococcus fascians, Heterodera zeae, Pratylenchus scribneri, Sonchus arvensis, Arabis Mosaic Virus, Impatiens Necrotic Spot Virus, Spodoptera littoralis, and Platynota stultana, the last two stand out.These AQPs had the second and fifth-highest scores on the AHP, evidencing the effectiveness of the method since these pests are widely discussed in the literature due to the high impact caused on their hosts (Van der Gaag and Van der Straten, 2017;Lázaro-Berenguer et al., 2022).
Although the countries of occurrence of Spodoptera littoralis and Platynota stultana are far from Brazil, which determined a lower score in the entry criterion, these pests are important due to their ability to spread.The main spread is through the transit of plants, which can be transported from one continent to another (Korycinska and Eyre, 2013).The pest S. littoralis has a flight capacity of approximately 50 km day-1 (Nasr et al., 1984), and the young larvae of P. stultana have the habit of ballooning in the wind on silk threads allowing floating in air currents.
Another point in common between S. littoralis and P stultana is the important impacts they can cause.S. littoralis is capable of reducing around 75% of cotton production (Espinosa and Hodges, 2009).Since Brazil has a cultivated area of 1,626,445 ha and a production value of U$ 5.33 billion (IBGE, 2020), losses can amount to U$ 2.4 billion.This result corroborates the description that S. littoralis is a polyphagous lepidopteran that is one of the most destructive crops in tropical and subtropical regions of the world (EPPO, 1997).Damage caused by P. stultana corresponds to approximately 50% production value of the attacked hosts.In grapes growing in Brazil, losses can reach U$ 334 million (Korycinska and Eyre, 2013).
The classification determined the low prioritization assigned by the method for Latheticus oryzae and Acarus siro composed group 1 and had lower results than the others, according to the AHP.Their final scores were low, mainly in the establishment and spread and estimated impacts criteria, because they are pests associated with stored grains, not affecting the hosts in the field (CABI, 2020).
The probability of entry of these pests, according to the AHP, is important.The occurrence of L. oryzae in Argentina (CABI, 2020) can facilitate its entry into the country due to the many border points considered vulnerable (Spadotto et al., 2014).Regarding A. siro, the method contains a high probability of entry.This entry potential was confirmed with its recent detection in Brazil (Barbosa et al., 2022).The economic impacts of A. siro are cited in stored grains of rice, corn, soybeans, wheat, and sorghum, in which it can reach 25% damage (Işikber et al., 2016;Clemmons and Taylor, 2016).Moreover, this species affects human health by causing allergies (Marques et al., 2022).
The probability of entry of these pests, according to the AHP, is important.The occurrence of Latheticus oryzae in Argentina (CABI, 2020) can facilitate its entry into the country due to the many border points considered vulnerable (Spadotto et al., 2014).Regarding A. siro, the method contains a high probability of entry.This entry criterion was confirmed with its recent detection in Brazil (Barbosa et al., 2022).The economic impacts of A. siro are cited in stored grains of rice, corn, soybeans, wheat, and sorghum, which can reach 25% damage (Işikber et al., 2016;Clemmons and Taylor, 2016).Moreover, this species affects human health, causing allergies (Marques et al., 2022).
The results obtained in the present study contain fundamental information for government decision-making in the development of public policies on absent quarantine pests.

Conclusions
The results obtained in the present study contain essential information for government decision-making in the development of public policies on absent quarantine pests.
The AHP method is effective for AQP prioritization for organisms of different biological groups since out of the 20 AQPs studied, 17 are the priority.This method provides substantial subsidies for developing research and phytosanitary defense strategies to prevent the introduction of AQPs in Brazil.
Other countries from different continents can use this methodology to prioritize AQPs and thus create strategic plans to prevent entry into their territories and economic, social, and environmental impacts.

Figure 1 .
Figure 1.Clustering of Absent Quarantine Pests (AQP) by the K-means algorithm, in the criteria of entry, establishment, and spread, and estimated impacts.

Table 1 .
List of Absent Quarantine Pests (AQP) selected according to the number of interceptions by Brazil and the European and Mediterranean Plant Protection Organization between 2015 and 2018, hosts, and countries in which their occurrence was recorded.NI, Number of interceptions, NH, Number of hosts, NC, number of countries.

Absent Quarantine Pests (AQP) Order: Family Common name Type of pest NI NH NC
References: Number of interceptions = interceptions carried out by the Ministry of Agriculture, Livestock, and Supply (MAPA) and by the countries that make up the European and Mediterranean Plant Protection Organization, Number of hosts = several pest hosts, Number of countries a = number of countries with the occurrence of the pest.

Table 2 .
List of criteria and sub-criteria used to determine the importance of twenty quarantine pests in Brazil.
CriterionSub-criteria Entry 1. Distance between the nearest location and the Brazilian border 2. Number of continents where the pest occurs 3. Number of imports of host material or regulated article 4. Number of countries in which the pest occurs 5. Number of countries bordering Brazil where the pest occurs 6. Import volume of host material or regulated article.15.Number of jobs in the host crop production chain 16.Number of properties with the host crop 17.Number of countries regulating the pest 18. Potential for contamination by pesticides 19.Value of the annual production of the host crop

Table 4 .
Ranking of absent quarantine pests (AQP), their prioritization (P), and score (S), according to the sub-criteria of the ESTABLISHMENT AND SPREAD criterion and subcriteria: 7, climate adaptation, 8, number of hosts, 9, the total area of host crops, 10, percentage of microregions with host crops, 11, efficiency of control methods, 12 annual natural spread distance estimation, 13, probability of human spread of the pest.

Table 5 .
Ranking of absent quarantine pests (AQP), their prioritization (P), and score (S), according to the sub-criteria of the Estimated impacts criterion.14, the expected percentage of damage, 15, the value of the annual production of the host crop, 16, number of countries regulating the pest, 17, number of properties with the host crthe op, 18, number of jobs in the host crop production chain, the 19, potential of contamination by pesticides.

Table 6 .
Ranking of absent quarantine pests (AQP), their prioritization (P), and score (S), according to the Entry (E), Establishment and spread (ES,) and estimated impacts (EI) criteria.

Table 7 .
Clustering of absent quarantine pests (AQP), according to the K-means algorithm.