Machine learning methods: Modeling net growth in the Atlantic Forest of Brazil

Ecological Informatics, Год журнала: 2024, Номер 81, С. 102564 - 102564

Опубликована: Март 18, 2024

Tree growth models are an important and essential part of modeling forest dynamics valuable tools for management planning biodiversity conservation strategies. We applied three different machine learning models, namely Artificial Neural Networks (ANN), Support Vector Machine (SVM) Random Forest (RF) to predict tree at the plot-level in Atlantic Brazil. attributes, land use history, landscape, soil climatic characteristics were used modeling. Recursive Feature Elimination was select best subset predictor variables. found that edaphic, attributes variables shaping Brazilian Forest. Soil acidity most characteristic. The methods efficient. method showed superiority over others Nemenyi test pointed out difference between RF model other techniques greater than calculated critical (CD). can be tool fragments They help understanding biome developing strategies aimed recovering reducing deleterious effects fragmentation.

Язык: Английский

---

Adaptation of SAVI to estimate the leaf area index considering different land covers in a Brazilian atlantic forest area

Modeling Earth Systems and Environment, Год журнала: 2025, Номер 11(2)

Опубликована: Янв. 23, 2025

Язык: Английский

---

Ecosystem resilience response to forest fragmentation in China: Thresholds identification

Journal of Environmental Management, Год журнала: 2025, Номер 380, С. 125180 - 125180

Опубликована: Апрель 1, 2025

Язык: Английский

---

Assessing Drought Vulnerability in the Brazilian Atlantic Forest Using High Frequency Data

Опубликована: Янв. 11, 2024

This research investigates the exposure of plant species to extreme drought events in Brazilian Atlantic Forest, employing an extensive dataset collected from 205 automatic weather stations across region. Meteorological indicators derived hourly data, encompassing precipitation, maximum and minimum air temperature, were utilized quantify past, current, future conditions. The dataset, comprising 10,299,236 data points, spans a substantial temporal window exhibits modest percentage missing data. Missing excluded analysis, aligning with decision refrain imputation methods due potential bias. Drought quantification involved computation Aridity Index, analysis consecutive hours without classification wet dry days per month. Mann-Kendall trend was applied assess trends evapotranspiration considering their significance. hazard assessment, incorporating environmental factors influencing tree growth dynamics, facilitated ranking meteorological identify regions most exposed events. results revealed consistent occurrences rainfall events, indicated by positive outliers monthly precipitation values. However, significant observed, including increase daily temperature coupled decrease Forest. No correlation between vulnerability ranks station latitudes elevation found, suggesting geographical location does not strongly influence observed dryness trends.

Язык: Английский

---

Assessing Drought Vulnerability in the Brazilian Atlantic Forest Using High-Frequency Data

Meteorology, Год журнала: 2024, Номер 3(3), С. 262 - 280

Опубликована: Июль 16, 2024

This research investigates the exposure of plant species to extreme drought events in Brazilian Atlantic Forest, employing an extensive dataset collected from 205 automatic weather stations across region. Meteorological indicators derived hourly data, encompassing precipitation and maximum minimum air temperature, were utilized quantify past, current, future conditions. The dataset, comprising 10,299,236 data points, spans a substantial temporal window exhibits modest percentage missing data. Missing excluded analysis, aligning with decision refrain using imputation methods due potential bias. Drought quantification involved computation aridity index, analysis consecutive hours without precipitation, classification wet dry days per month. Mann–Kendall trend was applied assess trends evapotranspiration considering their significance. hazard assessment, incorporating environmental factors influencing tree growth dynamics, facilitated ranking meteorological identify regions most exposed events. results revealed consistent occurrences rainfall events, indicated by positive outliers monthly values. However, significant observed, including increase daily temperature coupled decrease Forest. No correlation between vulnerability ranks station latitudes elevation found, suggesting that geographical location do not strongly influence observed dryness trends.

Язык: Английский

---

Limited influence from edges and topography on vegetation structure and diversity in Atlantic Forest

Plant Ecology, Год журнала: 2023, Номер 225(4), С. 361 - 371

Опубликована: Сен. 25, 2023

Язык: Английский

---

Limited influence from edges and topography on plant structural and taxonomic diversity in fragments of Atlantic Forest

Research Square (Research Square), Год журнала: 2023, Номер unknown

Опубликована: Июнь 5, 2023

Although Atlantic Forest is very diverse and heavily fragmented, little known about the impact of created edges on forest structure plant diversity within its remnants. We aimed to determine distance edge influence vegetation in fragments Forest; compare effects influence, topography their interaction structure; assess patterns structural taxonomic diversity. collected data structure, functional groups, families vertical 2 m x contiguous quadrats along 250 transects across 24 approx. 70 km west São Paulo. used randomization tests estimate magnitude generalized linear mixed model effect topography, wavelet analysis evaluate spatial patterns. found evidence degradation (lower cover most groups compared interior forest) sealing (abrupt changes at particularly for leafy diversity), but did not extend far into with a or less than 20 variables. Less extensive other tropical forests was explained by (slope) could be due more fragmentation land use history. The multiple approaches studying provided complementary information improve our understanding anthropogenic Forest.

Язык: Английский

---

Coarse woody debris dynamics in a secondary Atlantic Forest fragment in Brazil

Canadian Journal of Forest Research, Год журнала: 2024, Номер 54(6), С. 646 - 659

Опубликована: Янв. 19, 2024

The Atlantic Forest fragments have suffered from the impacts of climate change, resulting in increased production coarse woody debris (CWD), which needs to be evaluated space and time generate accurate estimates carbon accumulation. Thus, goals this study were (i) quantify CWD volume, necromass, stock, annual increment (AI carb ) over a period 4 years; (ii) select optimal combination climatic, topographic, edaphic, intrinsic forest variables accurately predict AI using machine learning multivariate analysis. stock between 2017 2020. was 1.09 MgC ha −1 year (2017–2018), 1.24 (2018–2019), 2.31 (2019–2020). Statistical analysis indicated that had greater weight 2018–2019 2019–2020 periods, while topographical, more important for 2017–2018 period. Our findings showed increase linked temporal spatial within forests. These results demonstrate importance parameter cycle ecosystems highlight there should international research efforts pool.

Язык: Английский

---

Development of Land Cover and Carbon Storage in Plawangan Hill of Gunung Merapi National Park, Yogyakarta, Using Landsat Data Series 2009, 2013, 2017, and 2023

Jurnal Manajemen Hutan Tropika (Journal of Tropical Forest Management), Год журнала: 2024, Номер 30(1), С. 107 - 117

Опубликована: Апрель 4, 2024

Globally, habitat loss, deforestation, and climate change are mostly caused by land cover changes (LCC). The amount of covered trees has had a major impact on global warming change. Increasing the helps to mitigate warming. This study aims investigate in carbon storage Plawangan Hill, Indonesia, over four years (2009, 2013, 2017, 2023). site was defined as conservation area that been periodically impacted both directly indirectly volcanic eruptions. Images from Landsat 7 8 were used collect data. Additionally, assessed using forest canopy density (FCD Mapper) model, which then utilized quantify research site. findings demonstrated fluctuations between 2009 2023. have direct storage. age trees, type vegetation, succession stage, history eruptions variables apparent be main causes these changes.

Язык: Английский

---

Assessing Drought Vulnerability in the Brazilian Atlantic Forest Using High Frequency Data

Опубликована: Июнь 20, 2024

This research investigates the exposure of plant species to extreme drought events in Brazilian Atlantic Forest, employing an extensive dataset collected from 205 automatic weather stations across region. Meteorological indicators derived hourly data, encompassing precipitation, maximum and minimum air temperature, were utilized quantify past, current, future conditions. The dataset, comprising 10,299,236 data points, spans a substantial temporal window exhibits modest percentage missing data. Missing excluded analysis, aligning with decision refrain imputation methods due potential bias. Drought quantification involved computation Aridity Index, analysis consecutive hours without classification wet dry days per month. Mann-Kendall trend was applied assess trends evapotranspiration considering their significance. hazard assessment, incorporating environmental factors influencing tree growth dynamics, facilitated ranking meteorological identify regions most exposed events. results revealed consistent occurrences rainfall events, indicated by positive outliers monthly precipitation values. However, significant observed, including increase daily temperature coupled decrease Forest. No correlation between vulnerability ranks station latitudes elevation found, suggesting geographical location does not strongly influence observed dryness trends.

Язык: Английский

---