Policy debates on habitat set-asides need local benefit estimates rather than imported meta-analytic means. We combine yield response functions from bagged versus open-pollinated tree rows with contingent valuation of hive rental among mango and citrus growers. Marginal pollination value per hectare exceeds typical pesticide budgets during peak bloom, yet willingness-to-pay for native hedgerow restoration rises only when extension agents demonstrate hive loss recovery after pesticide drift incidents.

Municipal plants increasingly discharge trace ibuprofen and carbamazepine even after secondary treatment. Microcosm experiments with three native sedge species exposed to spiked effluent show root uptake and partial metabolic transformation within fourteen days. Carex species accumulate the highest tissue concentrations yet maintain growth, whereas one riparian grass exhibits leaf necrosis at environmentally relevant mixtures, guiding species selection for polishing wetland cells downstream of conventional treatment.

Light-trap surveys under-sample taxa that avoid ultraviolet attractants. We deploy calibrated recorders along forest edge gradients and extract embedding vectors from short chirp segments using a pretrained audio network. Unsupervised clustering recovers seasonal turnover consistent with morphologically identified subsets, offering a scalable complement to traditional trapping when voucher specimens are limited by permit constraints.

Clinic referrals for child conduct issues rarely capture caregiver exhaustion as a concurrent pathway. We survey paired parents in dual-income households using validated burnout and sleep scales alongside teacher-rated behavior checklists. A mediation model indicates that sleep fragmentation partially transmits the association between burnout and externalizing scores, suggesting interventions targeting nighttime routines may yield dual benefits for caregiver wellbeing and child outcomes.

National soil inventories are sparse in montane grasslands where tillage history varies at fine scales. We fuse red-edge vegetation indices with kriging of archived plot measurements to produce wall-to-wall carbon stock maps. Cross-validation RMSE improves when topographic wetness is included as a secondary variable, yet uncertainty remains high on eroded slopes where spectral signals saturate, underscoring the need for stratified ground validation in degraded patches.

Post-discharge exercise programs often fail when daily logs feel burdensome. We compare a badge-based breathing exercise app with identical content delivered as printable worksheets among adults with moderate COPD. Twelve-week follow-up shows higher session completion and improved six-minute walk distance in the app arm, although benefits attenuate once push notifications are disabled, highlighting design choices that sustain engagement beyond novelty effects.

Torpor cycles may reshape microbial communities beyond simple abundance changes. We profile fecal metatranscriptomes from individually marked bats across pre-hibernation, mid-winter, and spring arousal phases. Carbohydrate fermentation pathways decline during deep torpor whereas urea cycle transcripts rebound briefly during interbout euthermy, suggesting microbial activity tracks host nitrogen recycling rather than remaining uniformly suppressed throughout winter dormancy.

Manual counting of spore traps is labor-intensive and prone to inter-observer drift during allergy season peaks. We train a lightweight U-Net on grayscale micrographs annotated by two mycologists and evaluate on independent slides from urban and forest sites. The model achieves high recall for ascospore morphotypes while flagging ambiguous clumps for human review, reducing analyst time per slide without missing threshold exceedances used in public health alerts.