Received: 23 October 2018 | Revised: 19 March 2019 | Accepted: 27 March 2019 DOI: 10.1111/jen.12643 ORIGINAL CONTRIBUTION Quantifying massive outbreaks of the defoliator moth Ormiscodes amphimone in deciduous Nothofagus‐dominated southern forests using remote sensing time series analysis Sergio A. Estay1,2 | Roberto O. Chávez3 | Ronald Rocco3 | Alvaro G. Gutiérrez4 1Instituto de Ciencias Ambientales y Evolutivas, Universidad Austral de Chile, Abstract Valdivia, Chile Forest insect outbreaks are one of the major biotic disturbances on natural and arti‐ 2 Center of Applied Ecology and ficial landscapes. Although abundant literature of insect outbreaks exists in the Sustainability, Pontificia Universidad Católica de Chile, Santiago, Chile Northern Hemisphere, studies for the Southern Hemisphere are rare. Recently, mas‐ 3Lab. Geo‐Información y Percepción sive outbreaks of the native moth Ormiscodes amphimone (Fabricius) (Lepidoptera: Remota, Pontificia Universidad Católica de Valparaíso, Instituto de Geografía, Hemileucinae) have been reported in the southern cone of South America. These Valparaíso, Chile O. amphimone outbreaks have defoliated large areas of temperate forests, raising 4 Departamento de Ciencias Ambientales y great concern among local inhabitants, but yet the spatio‐temporal patterns of these Recursos Naturales, Universidad de Chile, Santiago, Chile events have not been evaluated. Here, we quantify the extension of the massive O. amphimone outbreaks occurred in the Aysén region (southern Chile) in the period Correspondence Sergio A. Estay, Universidad Austral de 2000–2015 using a novel remote sensing approach and field data. Remote sensing Chile, Instituto de Ciencias Ambientales y detections were strongly in agreement with field observations and showed that mas‐ Evolutivas, Valdivia, Chile. Email: [email protected] sive outbreaks of O. amphimone are among the largest biotic disturbances in the for‐ ests of the Southern Hemisphere. Considering only field‐confirmed outbreaks, the Funding information Fondo Nacional de Desarrollo Científico defoliated area reached 164,000 hectares in total between 2000 and 2015. The es‐ y Tecnológico, Grant/Award Number: timation of the spatial impact of O. amphimone, and its recurrence, represents the 1160370; CONICYT PAI N, Grant/Award Number: 82140001; Fondecyt Iniciación, first step for the search of management alternatives of this massive disturbance. Grant/Award Number: 11171046 and 11150835; CAPES‐Conicy, Grant/Award KEYWORDS Number: FB‐0002; FIA PYT, Grant/Award defoliation, enhanced vegetation index, ormiscodes, remote sensing, southern beech forests Number: 2016‐0203 1 | INTRODUCTION of insect outbreaks as biotic perturbations on forests from New Zealand, Australia and the southern cone of South America (Hosking Despite most insects have little detectable impact on forests, some & Hutcheson, 1988; Loch & Floyd, 2001; Milligan, 1974; Paritsis & species may have detrimental effects on trees in terms of growth Veblen, 2011; Piper, Gundale, & Fajardo, 2015). rates, biomass productivity, carbon budget, age structure and spe‐ Recent scientific studies (Piper & Fajardo, 2014; Piper et al., cies composition and, in extreme cases, cause hosts' death (see 2015), technical reports from the Chilean Agricultural and Livestock Yang, 2012, for a review). Studies showing the impact of insect out‐ Service (SAG), the national plant protection organization of Chile breaks on natural forests are abundant in the Northern Hemisphere and local media have highlighted the occurrence of major insect out‐ (Haynes, Allstadt, & Klimetzek, 2014; Weed, Ayres, & Hicke, 2013). breaks of the native moth Ormiscodes amphimone causing complete In contrast, documented massive outbreaks are scarce in the natural defoliation of Nothofagus‐dominated temperate forests in South forests of the Southern Hemisphere. However, the few studies avail‐ America (46°S, Figure 1). The impressive blooming of O. amphimone able for this part of the world have highlighted the ecological impact caterpillars and concomitant forest defoliation have been observed J Appl Entomol. 2019;143:787–796. wileyonlinelibrary.com/journal/jen © 2019 Blackwell Verlag GmbH | 787 788 | ESTAY et AL. across thousands of hectares of forests, raising a great concern Despite this semi‐quantitative field evaluation of the defoliation among local inhabitants, tourist operators, local and national author‐ level, at the best of our knowledge, no studies have quantitatively ities. At eastern Patagonia, Paritsis, Veblen, Smith, and Holz (2011) evaluated the spatio‐temporal patterns of O. amphimone events in described a similar phenomenon, but apparently of a smaller spatial Chile. The large size of the broadleaf Nothofagus forests (Figure 1) scale. However, despite their apparent magnitude, the ecological and the inaccessibility of the affected areas are the main constraints mechanisms behind these events, their dynamics and the conse‐ for assessing the extension and impact of O. amphimone defoliations. quences of the outbreaks for the western Patagonia forests are still Unfortunately, the lack of an accurate localization and quantification unknown. of these outbreaks prevents the implementation of any management Ormiscodes amphimone (Lepidoptera: Saturniidae) is a native phy‐ strategy. tophagous moth (Figure 2). The larvae can reach 100 mm in length As an alternative to overcome these difficulties, over the past and 12 mm in width. Male adults range between 45 and 80 mm of two decades there has been a significant development of remote wingspan. On the other hand, female adults are slightly larger with sensing methods to assess insect‐related disturbances and their 58–95 mm wingspan (Angulo, Lemaire, & Olivares, 2004). Larvae feed consequences in forest ecosystems (e.g., Abdel‐Rahman, Mutanga, on several host plants such as Nothofagus spp., Populus spp., Prunus Adam, & Ismail, 2014; Eklundh, Johansson, & Solberg, 2009; spp., Juglands spp., Cryptocarya alba and Pinus spp., among many other Franklin, Fan, & Guo, 2008; Spruce et al., 2011). The ability to detect tree species (Angulo et al., 2004). However, major outbreaks are defoliation from satellite imagery depends on the type and degree of mainly described to start in Nothofagus pumilio forests in Chile and foliage damage, which may range from discoloration to total crown Argentina, but when the outbreak expands, larvae are able to defoli‐ defoliation and tree death (Eklundh et al., 2009). Also, it depends on ate almost any species present in the site. The species overwinters in the spatial extent of the outbreak in relation to the pixel size of the the egg stage. Larvae can be found feeding gregariously on host trees image. If the spots of defoliation are smaller or closer in size to the from August to January, and moth adults fly from January to June pixel size, then detection is not possible. Damage assessments are depending on location (Baldini & Alvarado, 2008; Paritsis & Veblen, commonly based on arithmetic combinations of reflectance bands 2011). In Chile, the species is distributed between 30° and 47°S that isolate the distinctive reflectance linked to a given leaf compo‐ (Angulo et al., 2004). Together with the negative effects on tourism nent (e.g., chlorophyll, leaf area index, water). The values obtained and outdoor activities due to its irritating hairs, O. amphimone is con‐ this way are the well‐known vegetation indices (Kerr & Ostrovsky, sidered detrimental for tree growth and timber production (Bauerle, 2003; Peña & Altmann, 2009). This technological approach to out‐ Rutherford, & Lanfranco, 1997), can kill saplings (Angulo et al., 2004; break quantification, although specially valuable for inaccessible re‐ Cogollor, Baldini, & Pancel, 2002) and potentially cause crown dieback gions like ours, requires a rigorous validation from field data. on Nothofagus spp. if defoliation is severe (Veblen, Hill, & Read, 1996). As a contribution to the evaluation of the impact of this insect on It is still not clear whether the outbreaks of O. amphimone follow southern forests and its management, in this study we evaluate the a cyclical or acyclical temporal pattern, and therefore, the disrup‐ spatial impact of massive O. amphimone disruptions in Nothofagus tion reported here cannot be considered yet as abnormal from its forests in South America in the time period 2000–2015 using historical natural behaviour. The few available reports seem to be novel remote sensing tools, followed by validation using field data. contradictory: Artigas (1972, 1994) reported some level of temporal Specifically, we used satellite imagery to (a) detect and geolocalize cyclicity, with outbreaks occurring every 4 years approximately at insect outbreaks from years 2000 to 2015 (b) estimate the level of ~30°S in Chile. On the other hand, a recent study of O. amphimone correspondence between the satellite‐derived and field data and (c) outbreaks on N. pumilio forests in the Argentinian Andes found no quantify the spatial extent of the outbreaks events. evidence of cyclic patterns (Paritsis & Veblen, 2011; Paritsis, Veblen, & Kitzberger, 2009). What seems to be a consensus for park rangers 2 | MATERIALS AND METHODS and local farmers is that outbreak events occur more frequently in the last two decades. 2.1 | Study area Despite multiple parasitoids have been described attacking Ormiscodes larvae and eggs, there is almost no information about Ormiscodes amphimone outbreaks have mainly been observed in the the population dynamics