Sub-alpine shrub classification using UAV images: Performance of human observers vs DL classifiers

Ecological Informatics, Journal Year: 2024, Volume and Issue: 80, P. 102462 - 102462

Published: Jan. 9, 2024

In recent years, the automatic analysis of natural environment images acquired with unmanned aerial vehicles (UAV) has rapidly gained popularity. UAVs are specially important in mountainous forests where access is difficult and large areas need to be surveyed. Zao mountains northeastern Japan, regenerated fir saplings competing sub-alpine vegetation shrubs after a severe tree mortality caused by bark beetle infestation. A detailed survey distribution key improve our understanding species succession influence climate change that process. To end, we evaluated suitability deep-learning-based image classification UAV order map potential regeneration. assess contribution this technology research field, first conducted an observer study difficulty for humans task classifying from images. Afterwards, compared observers' accuracy four state-of-the art deep learning networks classification. The best 55% demonstrates limitations using only Furthermore, sources error showed even though could differentiate between deciduous evergreen 96%, identifying correct within each group proved much more challenging. contrast, achieved values range 70–80% classification, clearly demonstrating capabilities beyond human experts. Our experiments also indicated performance these was significantly influenced similarity datasets used fine-tune them evaluate them. This fact highlights importance building publicly available databases further results. Nevertheless, results presented paper show UAV-acquired can usher new type large-scale study, spanning tenths or hundreds hectares high spatial resolution (of few cms per pixel), providing ability challenging dynamics problems go conventional fieldwork methodologies.

Language: Английский

---

Plankton dynamics across the freshwater, transitional and marine research sites of the LTER-Italy Network. Patterns, fluctuations, drivers

The Science of The Total Environment, Journal Year: 2018, Volume and Issue: 627, P. 373 - 387

Published: Feb. 3, 2018

Language: Английский

---

Functional composition and diversity of leaf traits in subalpine versus alpine vegetation in the Apennines

AoB Plants, Journal Year: 2020, Volume and Issue: 12(2)

Published: March 3, 2020

Mediterranean high mountain grasslands are shaped by climatic stress and understanding their functional adaptations can contribute to better ecosystems' response global change. The present work analyses the plant traits of high-elevation growing in limestone mountains explore, at community level, presence different strategies for resource use (conservative vs. acquisitive) diversity syndromes (convergent or divergent). Thus, we compared composition above-ground related acquisition subalpine alpine calcareous central Apennines, a region characterized dry-summer climate. We used georeferenced vegetation plots field-measured (plant maximum height, specific leaf area dry matter content) dominant species two characteristic types: Sesleria juncifolia Silene acaulis community. Both communities particular conservation concern rich endemic which measured here first time. analysed using community-weighted mean trait index Rao's function, assessed how much observed pattern deviated from random distribution calculating respective standardized effect sizes. results highlighted that an acquisitive strategy relatively higher prevail S. community, optimizing rapid carbon gain, would help overcome constraints exerted short season. divergent underlines co-occurrence grasslands, shows good adaptation microhabitat-rich environment. Conversely, grassland, conservative lower likely level resistance aridity over longer Our outcomes indicate preadaptation grassland shift upwards zone will become warmer drier as result anthropogenic climate

Language: Английский

---

Global change on the roof of the world: Vulnerability of Himalayan otter species to land use and climate alterations

Diversity and Distributions, Journal Year: 2021, Volume and Issue: 28(8), P. 1635 - 1649

Published: July 20, 2021

Abstract Aim Climate Change Vulnerability Assessment (CCVA) prescribes the quantification of species vulnerability based on three components: sensitivity, adaptive capacity and exposure. Such assessments should be performed through combined approaches that integrate trait‐based elements (e.g., measures sensitivity such as niche width) with correlative tools quantifying exposure (magnitude changes in climate within habitat). Furthermore, land use alterations may increase impacts biodiversity, CCVAs focus both change effects. Unfortunately, most have so far focused exclusively to change. Location Himalaya. Methods We evaluated otter occurring Himalayan region, is, Aonyx cinereus, Lutra lutra Lutrogale perspicillata , 2050 recently proposed Niche Factor Analysis (CNFA) framework Species Distribution Models. Results Future will reduce (6%–15%) shift (10%–18%) geographical range Himalaya, exerting more severe effects than Among components, played a greater role determining otters. Specifically, specialist species, L. perspicillata, showed highest comparison generalist, . Main conclusions Our results underline how coupling components can generate diverging predictions compared relying only. Moreover, intrinsic proved significantly important extrinsic metrics habitat

Language: Английский

---

Editorial: Mountain landscapes: Protected areas, ecosystem services, and future challenges

Ecosystem Services, Journal Year: 2021, Volume and Issue: 49, P. 101302 - 101302

Published: May 16, 2021

Mountain regions provide crucial ecosystem services (ES) to their inhabitants and the surrounding lowland populations. However, multiple anthropogenic drivers, such as land-use change, climate overexploitation, population growth severely affect mountain socio-ecological systems, with huge effects on ES provision. This Special Issue (SI) collects 29 contributions addressing (i) current research gaps in our understanding of landscape dynamics ES, (ii) protected areas, (iii) new approaches a holistic assessment systems. introduction SI summarizes major findings related impacts global change role challenges requirements develop sustainable management strategies for landscapes light future challenges. The emphasize specific environmental cultural characteristics well importance large part population. Due regional differences characteristics, needs values local communities need be better integrated into decision-making, supported by inter- trans-disciplinary cooperation, which would also lead improved provision at level.

Language: Английский

---