Examining the utility of hyperspectral remote sensing and partial least squares to predict plant stress responses to sulphur dioxide pollution: a case study of Trichilia dregeana Sond.

Abstract

The use of air quality monitoring stations is expensive, with pollution data being either unavailable or inaccessible. Hence, effects of atmospheric sulphur dioxide (SO2) levels on biomarkers related to environmental stress were investigated for Trichilia dregeana tree leaves, in order to assess their bioindicator potential. Leaves were sampled randomly from trees at three industrial sites within the South Durban Basin, and an ex situ control, across two seasons (n = 28, per season). Ground-level SO2 concentrations were measured daily and ranged between 1 and 25 ppb. There were significant (p < 0.001) differences across sites and seasons for leaf area and leaf chlorophyll content. Partial least squares regression (PLSR) was used to quantify the relationship between biomarkers and hyperspectral data. For leaf chlorophyll content and leaf area, r2 values ranged from 0.325–0.475 to 0.429–0.586, with root mean square error of prediction (RMSEP) ranging between 8.75–8.98 and 9.20–12.52. The variable importance in projection (VIP) method was utilized and significant hyperspectral wavebands were identified, within the red-edge region, at 552 and 704 nm for spring, and at 552 and 708 nm for summer. Notably, PLSR was able to relate hyperspectral data-sets to both biomarkers, showing promise in identifying stress in T. dregeana leaves. However, the interaction between leaf chlorophyll content and leaf area suggests that a simultaneous prediction of these biomarkers may be more suitable.