Abstract:How messages are framed shapes what audiences remember days later. In a preregistered online experiment, participants read short vignettes emphasizing safety, community benefit, or scientific consensus. Delayed free recall favors consensus-oriented frames when numeric side-effect rates are also provided. We discuss ethical boundaries and implications for campaign design beyond headline testing.
Abstract:Volunteer observations can scale mortality mapping if bias is understood. We link geotagged mobile reports with annual orthomosaic color indices across a drought-impacted watershed. Agreement improves when training filters remove duplicate submissions near access roads, yielding a hybrid layer that catches scattered mortality missed by spectral thresholds alone.
Abstract:Over-drying raises checking risk while under-drying delays planing schedules. We embed low-cost dielectric probes at multiple heights within industrial stickered stacks and compare inferred moisture profiles with gravimetric cores. Online profiles flag reversed gradients during fan reversals, enabling mid-cycle set-point adjustments that reduce final moisture spread without extending total kiln time.
Abstract:Narrow woody buffers are promoted to intercept runoff yet their pollinator habitat value is rarely quantified. We record timed flower visits across buffer widths and native shrub mixes bordering cereal fields. Network modularity increases with floral continuity but not buffer width beyond a threshold, indicating design rules should emphasize bloom succession rather than maximal planting depth alone.
Abstract:Ground surveys lag behind canopy discoloration visible from orbit. We train gradient-boosted classifiers on harmonized Sentinel-2 composites augmented with topographic wetness and wind-exposure indices. Holdout landscapes from two outbreak years achieve usable precision at patch scale ten to fourteen days before confirmed trap catches, suggesting prioritization maps for sanitation felling crews.
Abstract:Organic amendments are applied to jump-start biological activity on compacted skid trails, yet community assembly pathways remain unclear. We profile ITS amplicons across a factorial field trial and construct co-occurrence networks at three sampling depths. Amendment increases hub taxa associated with lignocellulose decomposition while preserving ectomycorrhizal modules linked to seedling survival in the second growing season.
Abstract:Street-tree programs are evaluated by survival counts yet shade delivery varies with crown architecture. We combine hemispherical photography with wearable wet-bulb globe temperature logging along four boulevard typologies. Gap fraction explains more variance in afternoon exposure than stem density alone, supporting species mixes that prioritize broad crowns over fast juvenile height growth.
Abstract:Ring-width series from long-term monitoring plots often mix age trends with climate signals. We develop a mixed-effects detrending workflow that preserves high-frequency latewood variation linked to standardized precipitation indices. Cross-dated cores from three thinning regimes show divergent sensitivity after canopy closure, informing drought-oriented silvicultural prescriptions without discarding legacy chronologies.
Abstract:Residual active pharmaceutical ingredients challenge conventional biological treatment. We compare suspended and immobilized photocatalysts under simulated solar irradiation for representative analgesic and antibiotic compounds. Matrix effects from organic matter slow apparent rate constants; pre-oxidation modestly restores performance. Lifecycle notes emphasize catalyst recovery and energy use relative to ozonation baselines.
Abstract:Independent disruption models underestimate tail risk when suppliers face common environmental or logistics shocks. We propose scalar resilience indices based on time-to-recover distributions under a copula-linked hazard model. Case illustrations use stylized multi-tier networks. Sensitivity analysis shows that diversification yields diminishing returns when correlation exceeds moderate levels, motivating targeted redundancy instead of uniform splitting.