Hyperspectral evidence of early-stage pine shoot beetle attack in Yunnan pine

Highlights

• Manipulated infestation revealed the crown spectral signature under pine shoot beetle attack.

• 4 weeks’ pine shoot beetle shoot-feeding attack turned most shoots from green to red.

• Crown heterogenous discoloration resulted in a decline of chlorophyll and water content.

• Shoot damage influenced the retrieval accuracy of physiological parameter via hyperspectral data.

• 35 spectral variables were evaluated in random forest for identifying shoot damage ratio.

Abstract

Pine shoot beetle (PSB, Tomicus spp.) outbreaks cause widespread Yunnan pine (Pinus yunnanensis Franch) mortality in southwestern China. Early identification of PSB attacks could help forest managers mitigate the infestation before it turns into an outbreak. However, the subtle spectral changes and complex process of PSB-induced crown discoloration make the remote sensing approach difficult.

This study employed a manipulated insect infestation experiment to reveal suitable monitoring indicators. Healthy Yunnan pine crowns were infested with PSB adults at different preset densities (light, moderate, and severe) using the bagging method. The crown damage parameters, physiological properties, and corresponding spectral data were systematically measured in time series. Partial least squares regression (PLSR) was used to retrieve crown chlorophyll_a+b (Cab) and relative water content (RWC) via band reflectance. Random forest (RF) was used to determine the optimum spectral variables capable of capturing dynamic variations in crown shoot damage ratio (SDR).

Results showed that (1) after four weeks’ PSB shoot feeding attack, SDR reached saturated (26%-50%) for all damage levels, indicating a crown discoloration “switch point”. (2) A continuous decline was found in Cab and RWC at both shoot and crown levels during the whole infestation process, and the crown level decrease was smaller than the shoot level due to crown heterogeneous discoloration. (3) For PLSR retrieving crown Cab and RWC, estimation accuracy of healthy crowns could reach R² = 0.70, RMSE_cv = 14.60 mg/m² for Cab, and R² = 0.82, RMSE_cv = 1.25% for RWC, respectively. However, PSB early attack impacted these retrieval accuracies. (4) Spectral differences were evident in the visible region (530–600 nm) and near-infrared (NIR) plateau (780–1300 nm) among the preset damage levels. Significant differences in spectral variables (AMP, CI, OSAVI, RD, and SR) were observed between healthy and moderately damaged crowns. (5) In RF modeling, the importance of spectral variables for SDR estimation varied throughout the shoot feeding process. OSAVI, CI, MSI, PSRI, and SR2 were the top indicators that were suitable for identifying SDR ranging from 5% to 50%. SDR could be estimated with an accuracy of R² = 0.71, RMSE_cv = 8.92%, after 4 weeks’ infestation. Simultaneously, with the RF model, 4 preset damage levels could be differentiated with an out-of-bag error of 14.94% and a kappa coefficient of 0.8835.

In conclusion, this study mainly examined the physiological and spectral signatures of PSB early attacked crowns, and provided optimum spectral variables and models which will further help unmanned aerial vehicle (UAV) and satellite remote sensing identify PSB infestation at the early stages.

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.