Hyperspectral evidence of early-stage pine shoot beetle attack in Yunnan pine
Introduction
Disturbances by insects are natural processes in forest ecosystems and an integral driver of their dynamics, helping to maintain healthy and heterogeneous forests that can provide important ecosystem services (Raffa et al., 2009). However, many forest ecosystems have experienced an increase in the rate, magnitude, and frequency of insect disturbances, with recent disturbance activity considerably exceeding levels known from 20th-century experience (Millar and Stephenson, 2015). In China, forest pest outbreaks have been in a high incidence for a long time, causing an annual loss of 5 billion yuan. Among these pests, the pine shoot beetle (PSB) is one of the representatives and is considered to be the most aggressive one in southwestern China. Over the past 40 years, twice PSB outbreaks impacted over 1.5 million ha (destroyed 0.2 million ha) of Yunnan pine (Pinus yunnanensis Franch), significantly affecting the sustainable development of local forests (Lieutier et al., 2015). Therefore, it is of great importance to find an effective way to identify the location and severity of the PSB attacks for the purpose of minimizing the economic and ecological damage to forest health.
PSB (Tomicus spp.) was first described in Linnaeus’ magnum opus Systema naturae as follows: “lives in younger pine shoots, which it hollows out, dries out, hence acting as nature’s gardener in this tree” (Linnaeus, 1758), showing that Linnaeus was well aware of its shoot feeding habit and corresponding influence on tree crown appearance. The life history of PSB is complex and still not completely clear. However, its univoltine lifecycle could be roughly divided into two phases: the shoot-feeding phase and the trunk-colonizing phase. At the shoot-feeding phase, adult beetles feed on healthy shoots of tree crowns, absorb nutrients, and achieve sexual maturity. As shoot feeding attack causes substantial tree growth loss (Långström and Hellqvist, 1991, Eidmann, 1992, Czokajlo et al., 1997), reduces tree resistance, and predisposes these trees to subsequent trunk colonization (Schlyter and Löfqvist, 1990; Långström and Hellqvist, 1993; Lieutier et al., 2003), this phase is believed to be the early stage of attack. At the trunk-colonizing phase, the beetles construct maternal galleries for mating and reproduction in the trunk’s inner bark, resulting in visible red crowns (Ye and Li, 1994; Långstrӧm et al., 2002). For integrated pest management (IPM) of this pest, cutting down the colonized trees (red crown) alone cannot effectively control the PSB population (Zhang, 2002). By the time the colonized trunk is located via field investigation, the beetles would have long abandoned the tree and initiated new infestations, contributing to the next outbreak. Therefore, these factors altogether make the accurate early monitoring technology of PSB shoot feeding attack a crucial step in its disaster management.
Crown shoot damage ratio (SDR) is a commonly used indicator to quantify the severity of PSB attack (Standard of Forest Pest Occurrence and Disaster; LYT1681-2006). Field investigation and expert grading are followed to measure this indicator. However, due to the high cost involved in the ground survey and insufficient observation scale, it is almost impossible to repeatedly obtain standardized information from large areas with high frequency (Pause et al., 2016). While in-situ terrestrial forest monitoring is generally applied on a plot level, remote sensing approaches provide wall-to-wall information on multiple temporal and spatial scales (Lausch et al., 2016). Particularly, hyperspectral remote sensing provides an excellent chance to monitor plant early stress caused by forest pests (Senf et al., 2017, Stone and Mohammed, 2017, Hernández-Clemente et al., 2019). Based on narrow-band reflectance and fine spatial resolutions, hyperspectral data can retrieve the structural and physiological properties, such as plant chlorophyll, water content, and environmental stress level (Underwood et al., 2003, Carlson et al., 2007, Kokaly et al., 2009, Calderón et al., 2013, Thenkabail et al., 2016), making it an ideal approach to identify early-stage PSB attack.
Several attempts at different scales (leaf, crown, and plot) have been made to document the effect of PSB damage on the tree's physiological and spectral characteristics. Wang et al. (2018) analyzed the spectral features of pine damaged shoots and set up model judgment rules for shoot damage level classification. Lin et al. (2018) found that the radiative transfer model (RTM) LIBERTY improved the simulation accuracy of heterogeneous needle reflectance, making it suitable to assess PSB attacks. Liu et al. (2020) investigated the correlation between the damaged needles’ photosynthetic activity and hyperspectral data and set up photosynthetic parameter simulation equations. Wang et al. (2019) studied the relationship between needle temperature and the damaged shoots’ physiological status and then explored the feasibility of thermal infrared images for PSB damage monitoring.
Despite all these efforts, several critical knowledge gaps still exist, making it difficult for remote sensing to identify PSB early attack. Firstly, PSB-induced crown discoloration has a lag of several days since the initial attack. As the shoot-feeding process is gradual and complex, the suitable time window for remote sensing is unclear. Besides, PSB bores into the shoots and block water and nutrient delivery to the needles. This leads to the needle color of the damaged shoots changing from green to red during the infestation stages, while those undamaged shoots remain green. PSB thereby causes heterogeneous needles with patchy chlorosis which strongly influences the whole leaf and canopy reflectance due to the non-uniform distribution of chlorophyll. Compared with other bark beetles’ damage symptoms (e.g. red attack and grey attack) (Coops et al., 2006, Bright et al., 2013), this partial crown dieback symptom caused by PSB shoot feeding damage is less visually distinct than the uniformly (systematically) discolored red crown. Although studies of spectral variables on leaf scale damage have been conducted (Wang et al., 2019, Liu et al., 2020), crown scale physiological traits and optimal spectral variables have not been documented. Further research should reveal the complex mechanistic links among these factors.
Therefore, in order to develop a useful remote sensing approach for identifying PSB early attack and thereby help successfully manage the outbreaks of this pest, we carried out a manipulated insect infestation experiment by artificially infesting PSB adults on Yunnan pine crowns. Under this controlled condition, ground-based hyperspectral and physiological measurements were taken simultaneously to achieve the following objectives: (1) record the time series of early attacked crown spectral signatures and physiological properties under different PSB attack densities. (2) analyze the potential correlation between PSB attack severity, crown damage parameters and physiological parameters, and examine the impact of early-stage PSB attack on the crown physiological retrieval with hyperspectral measurements. (3) evaluate the performance of spectral variables in estimating SDR based on machine learning algorithm (i.e. random forest, RF); and (4) establish a model approach of early-stage PSB damage (crown SDR / damage level) estimation. Our findings will provide the optimal spectral variables and models for UAV- and satellite-based remote sensing in PSB pest disaster monitoring and early warning systems.
Section snippets
Manipulated infestation experiment
During the PSB shoot feeding phase in July 2018, a manipulated infestation experiment was performed in a five-year-old Yunnan pine afforestation of Qujing City, Yunnan Province (103°47′E, 25°39′N, 2000 m a.s.l.; Fig. 1a). Healthy Yunnan pine trees with similar size and growth were selected before the infestation so that significant variation in initial physiological and structural signatures among preset treatments would be avoided. Diameter-at-breast-height (DBH), tree height, and crown size
Time series of crown damage at the early stages of PSB attack
The newly attacked trees showed gradual crown discoloration and reddish shoot dieback (Fig. 3a). This appearance took approximately two weeks since the initial attack. After that, green, yellow, and red shoots were observed simultaneously within the same crown. The crown damage parameters (YSR, RSR, and SDR) increased with prolonged infestation and enhanced damage level (intense shoot feeding) (Fig. 3b).
Initially, the YSRs of all the damage levels were above 10% within one week while RSR was
Influence of manipulated infestation experiment on the discoloration process
The previous study assumed that Yunnan pines with typical symptoms of different damage levels are at various stages of attack (Liu et al., 2019), which follows the idea of “replacing time with space” and ignores the dynamic changes that damaged Yunnan pines undergo. Therefore, a novel manipulated infestation experiment was adopted to simulate the whole PSB shoot-feeding process, so that single crown level hyperspectral and physiological measurements could be taken in time series, and suitable
Conclusions
There are rising concerns regarding the impact of insect disturbances on biogeochemical cycling (Edburg et al., 2012), including carbon cycle (Kurz et al., 2008, Seidl et al., 2014), biodiversity (Beudert et al., 2015, Müller et al., 2008), and forest economic value (Dale et al., 2001). PSB outbreak can be devastating, with consequences ranging from widespread mortality of Yunnan pine to loss of slope stability and changes in carbon, water, and energy balance of the plantations (Lü et al., 2014
Author contributions
Yujie Liu, Lili Ren, and Youqing Luo designed this experiment; Yujie Liu carried out the fieldwork, analyzed the results, and wrote the manuscript; Sangzi Ze, Zhongyi Zhan and Linfeng Yu facilitated the fieldwork. We are grateful to the editor and the reviewers for their insightful comments and highly constructive criticisms that helped improve the manuscript’s content and presentation.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This study was supported by (1) the National Key Research & Development Program of China, “Research on key technologies for prevention and control of major disasters in plantation” (2018YFD0600200) and (2) Major Emergency Science and Technology Project of National Forestry and Grassland Administration (ZD202001).
References (95)
- et al.
European spruce bark beetle (Ips typographus, L.) green attack affects foliar reflectance and biochemical properties
Int. J. Appl. Earth Obs. Geoinf.
(2018) Biophysical and biochemical sources of variability in canopy reflectance
Remote Sens. Environ.
(1998)- et al.
High-resolution airborne hyperspectral and thermal imagery for early detection of Verticillium wilt of olive using fluorescence, temperature, and narrow-band spectral indices
Remote Sens. Environ.
(2013) - et al.
Hyperspectral discrimination of tropical rain forest tree species at leaf to crown scales
Remote Sens. Environ.
(2005) - et al.
Assessment of QuickBird high spatial resolution imagery to detect red attack damage due to mountain pine beetle infestation
Remote Sens. Environ.
(2006) Remote sensing of foliar chemistry
Remote Sens. Environ.
(1989)- et al.
Estimating corn leaf chlorophyll concentration from leaf and canopy reflectance
Remote Sens. Environ.
(2000) - et al.
High resolution derivative spectra in remote sensing
Remote Sens. Environ.
(1990) - et al.
PROSPECT-4 and 5: Advances in the leaf optical properties model separating photosynthetic pigments
Remote Sens. Environ.
(2008) - et al.
A Narrow-waveband spectral index that tracks diurnal changes in photosynthetic efficiency
Remote Sens. Environ.
(1992)
NDWI - A Normalized Difference Water Index for Remote Sensing of Vegetation Liquid Water from Space
Proc. SPIE – Int. Soc. Optical Eng..
Simulation of satellite, airborne and terrestrial LiDAR with DART (I): Waveform simulation with quasi-Monte Carlo ray tracing
Remote Sens. Environ.
Partial least-squares regression: a tutorial
Anal. Chim. Acta
The Chlorophyll Fluorescence Ratio F735/F700 as an Accurate Measure of Chlorophyll Content in Plants
Remote Sens. Environ.
Use of a green channel in remote sensing of global vegetation from EOS-MODIS
Remote Sens. Environ.
Integrated narrow-band vegetation indices for prediction of crop chlorophyll content for application to precision agriculture
Remote Sens. Environ.
A comparison of vegetation indices global set of TM images for EOS–MODIS
Remote Sens. Environ.
A comparison of regression tree ensembles: predicting Sirex noctilio induced water stress in Pinus patula forests of KwaZulu-Natal, South Africa
Int. J. Appl. Earth Obs. Geoinf.
Characterizing canopy biochemistry from imaging spectroscopy and its application to ecosystem studies
Remote Sens. Environ.
Random Forest ensembles for detection and prediction of Alzheimer’s disease with a good between-cohort robustness
Neuroimage Clin.
An investigation into robust spectral indices for leaf chlorophyll estimation
ISPRS J. Photogramm. Remote Sens.
Nondestructive measurement of fruit and vegetable quality by means of NIR spectroscopy: a review
Postharvest Biol. Technol.
Reflectance indices associated with physiological changes in nitrogen and water limited sunflower leaves
Remote Sens. Environ.
Optimization of soil-adjusted vegetation indices
Remote Sens. Environ.
Remote sensing of forest insect disturbances: Current state and future directions
Int. J. Appl. Earth Obs. Geoinf.
Use of hyperspectral derivative ratios in the red-edge region to identify plant stress responses to gas leaks
Remote Sens. Environ.
Red and photographic infrared linear combinations for monitoring vegetation
Remote Sens. Environ.
Mapping nonnative plants using hyperspectral imagery
Remote Sens. Environ.
Monte carlo cross validation
Chemom. Intell. Lab. Syst.
Leaf chlorophyll content retrieval from airborne hyperspectral remote sensing imagery
Remote Sens. Environ.
Bark beetles increase biodiversity while maintaining drinking water quality
Conser. Lett.
Random forests
Mach. Learn.
Predicting live and dead tree basal area of bark beetle affected forests from discrete-return lidar
Can. J. Remote Sens.
Variability and application of the chlorophyll fluorescence emission ratio red/far-red of leaves
Photosynthesis Res
Hyperspectral remote sensing of canopy biodiversity in Hawaiian lowland rainforests
Ecosystems
Partial least squares regression as an alternative to current regression methods used in ecology
Oikos
Ratios of leaf reflectance in narrow wavebands as indicator of plant stress
Int. J. Remote Sens.
Leaf optical properties in higher plants: linking spectral characteristics to stress and chlorophyll concentration
Am. J. Bot.
Weighted Kappa: nominal scale agreement with proVIion for scaled disagreement or partial credit
Psychol. Bull.
Remote sensing of crop type and maturity
Photogramm. Eng. Remote Sens.
Growth reduction of Scots pine, Pinus sylvestris, caused by the larger pine shoot beetle, Tomicus piniperda (Coleoptera, Scolytidae), in New York State
Can. J. For. Res.
Climate change and forest disturbances
Bioscience
Cascading impacts of bark beetle-caused tree mortality on coupled biogeophysical and biogeochemical processes
Front. Ecol. Environ.
Impact of bark beetles on forests and forestry in Sweden
J. Appl. Entomol.
The red-edge position and shape as indicators of plant chlorophyll content, biomass and hydric status
Int. J. Remote Sens.
Remote estimation of chlorophyll content in higher plant leaves
Int. J. Remote Sens.
Cited by (14)
Using only the red-edge bands is sufficient to detect tree stress: A case study on the early detection of PWD using hyperspectral drone images
2024, Computers and Electronics in AgricultureEarly detection of pine shoot beetle attack using vertical profile of plant traits through UAV-based hyperspectral, thermal, and lidar data fusion
2023, International Journal of Applied Earth Observation and GeoinformationClassification of pine wilt disease at different infection stages by diagnostic hyperspectral bands
2022, Ecological IndicatorsCitation Excerpt :There is a significant difference in moisture content between healthy and early stages, so the related spectral parameters have good potential for early identification. This result was consistent with the research of other stem borers(Abdullah et al., 2018; Liu et al., 2021). The absorption spectrum around the green peak and the SWIR bands reflected the decreased water and chlorophyll content (Jones and Vaughan, 2010; Roberts et al., 2012).
Automated detection of an insect-induced keystone vegetation phenotype using airborne LiDAR
2024, Methods in Ecology and Evolution