Ecological sensitivity and urban risk screening of aquatic and terrestrial species exposed to pre- and post-emergent herbicides

Herbicides applied to agricultural soils can adversely affect non-target organisms in both terrestrial ecosystems and downstream aquatic environments. However, comparative evaluations of ecological sensitivity across herbicide classes and environmental compartments remain limited. This study assessed five pre-emergent (PRE) herbicides (simazine, atrazine, trifluralin, diuron, and pendimethalin) and five post-emergent (POST) herbicides (glyphosate, 2,4-D, paraquat, glufosinate-ammonium, and dicamba) using species sensitivity distributions (SSDs) and evaluated whether the structural descriptor LogP could explain observed ecological sensitivity patterns in soil and aquatic systems. Toxicity datasets were harmonised across five representative taxa per compound, encompassing primary producers and higher trophic-level organisms. Aquatic taxa included diatoms, green algae, duckweed, crustaceans, and fish, whereas terrestrial taxa comprised earthworms, springtails, woodlice, snails, and monocotyledonous plants, thereby enhancing ecological representativeness. Hazardous concentrations affecting 10% of species (HC10) derived from SSDs were subsequently used to calculate predicted no-effect concentrations (PNECs). SSDs revealed consistent differences in ecological sensitivity between herbicide classes. In aquatic environments, PRE herbicides exhibited substantially lower PNECs (0.0004–0.034 mg L −1 ) compared with POST herbicides (0.041–2.35 mg L −1 ), representing approximately one to two orders of magnitude difference in protective thresholds. Increased sensitivity was primarily associated with aquatic primary producers, particularly diatoms and macrophytes, which defined the lower tails of the SSDs. In soil systems, PNECs were generally higher for both herbicide groups; however, PRE herbicides consistently exhibited lower minimum threshold values than POST herbicides. Additionally, SSDs for PRE herbicides displayed steeper slopes, indicating reduced margins between no-effect concentrations and community-level responses. Exploratory QSAR analyses demonstrated limited predictive capacity of lipophilicity and related physicochemical descriptors, particularly within herbicide classes and for terrestrial taxa. These findings indicate that SSD-derived PNECs more effectively capture ecologically relevant sensitivity patterns than QSAR-based predictions alone, emphasising the importance of differentiating between PRE and POST herbicides in ecological risk assessment frameworks. • SSDs reveal higher aquatic sensitivity of pre-emergent than post-emergent herbicides. • Aquatic PNECs for pre-emergent herbicides are 1–2 orders of magnitude lower. • Primary producers dominate lower SSD tails for pre-emergent herbicides. • QSAR descriptors alone poorly predict ecological sensitivity across compartments. • SSD-derived PNECs support comparative screening of urban herbicide risks.