The importance of land use and cover change (LUCC) has gradually attracted more attention due to its influence on the climate and ecosystem. Consequently, the necessity of accurate LUCC mapping has ...become increasingly apparent. Over the past decades, although a large number of machine learning algorithms have been developed to improve the accuracy and reliability of remote sensing image classification, especially for LUCC classification, there is a lack of studies that assess the performance of machine learning algorithms in arid desert-oasis mosaic landscapes. In this study, the main objective is to provide a reference for the extraction of LUCC information in dryland regions with oasis-desert mosaic landscapes by comparing the performances of the k-nearest neighbor (KNN), random forest (RF), support vector machine (SVM) and artificial neural network (ANN) for the LUCC classification of the Dengkou Oasis, China. Landsat-8 Operational Land Imager (OLI) image data were used with spectral indices and auxiliary variables that were derived from a digital terrain model to classify 7 different land cover categories. The highest overall accuracy was produced by the ANN (97.16%), which was closely followed by the RF (96.92%), SVM (96.20%), and finally KNN (93.98%); statistically similar accuracies were obtained for the ANN, SVM and RF. The RF algorithm performed well across several aspects, such as stability, ease of use and processing time during the parameter tuning. Overall, the random forest algorithm is a good first choice method for land-cover classification in this study area, and the elevation and some spectral indices, such as the NDVI, MSAVI2 and MNDWI, should be used as variables to improve the overall accuracy.
Display omitted
•The performance of four machine learning algorithms for land use/cover classification over Dengkou Oasis was assessed.•The best classification accuracies were achieved by the artificial neural network and random forest algorithm.•Random forest algorithm performed well in stability, ease of use and processing time during the parameter tuning.•Random forest algorithm is a good first choice method for land use/coverclassification in Dengkou Oasis.
Evaluating how land cover is being transformed is essential to identify patterns necessary to infer the change trajectories and the driving factors. This study considers the case of Nigeria, where ...various natural ecosystems are being converted and for which a current national scale assessment at high spatial resolution is lacking. Producing 30 m Landsat-based time-series data, we analyze change among land cover types (i.e. tree-covered area, grassland, wetland, waterbody, cropland, artificial surface, and otherland) across seven agroecological zones. The annual change intensity was assessed at multi-levels across two time-intervals (i.e. 2000–2013, 2013–2022). Distinguishing between natural land cover and human activity-related land-use, we estimate the extent of change signifying how humans have appropriated natural land cover. Insights from analysis at the interval level reveal that land transformation accelerated from 3.3% in 2000–2013 to 4.5% during 2013–2022 in all agroecological zones (e.g. rainforest, mangrove), except in Sudan savannah and Sahel savannah where speed was higher in 2000–2013 as grasslands were increasingly cultivated. Cropland expanded almost two-fold (22% to 37%), whereas tree-cover declined from 50% to 31% and wetland from 7% to 3.7% over the 23 years. Much loss of natural land cover (e.g. tree-cover, grassland, and wetland) to cropland occurred in 2000–2013 (22%) when most irrigation schemes in Nigeria were established. In contrast, the loss of natural land cover to settlement (0.9%) during 2000–2013 increased to 2.0% in 2013–2022. Of all agroecological zones, the mangrove zone was most disturbed as its persisting land cover areas reduced from 69% to 5% between 2000–2013 and 2013–2022. The amount of persisting land cover was highest in the Sudan savannah at 44% in 2000–2013 and 49% in 2013–2022. Processes of human-appropriated natural land cover in Nigeria are related to urbanization and cropland expansion into natural areas with some instances of natural regeneration, especially in croplands and abandoned settlement areas.
There is a growing demand for incorporating social preferences in ecosystem service assessments. To establish sustainable rural landscape management with alternative sources of preferable ecosystem ...services that fulfills conservation objectives and reduces poverty, we need information about the perceptions of local people regarding particular landscape elements as the sources. Within a forest to agricultural landscape in West Java, we examined what types of ecosystem services were perceived by rural people, the effects of socioeconomic factors on people׳s perceptions of ecosystem services, and which landscape elements were perceived as the sources. We found rural people were greatly aware of ecosystem services, although more provisioning services were perceived than other services. Place of origin, residential location, area of agricultural lands and agroforests, and number of livestock were the most influential socioeconomic factors determining the number of ecosystem services an individual respondent perceived. People living closer to the remnant forest perceived more ecosystem services. Agroforest was perceived as the source providing the most ecosystem services followed by the forest. Our findings suggest people living close to forest will promote prospective ecosystem services as long as their needs are accommodated. Furthermore, our findings indicate agroforests play an important role as a supplement of forests.
•We examined perceptions of local people regarding ecosystem services (ESs).•Local people living close to the remnant forest appreciated many ESs.•Agroforests were seen as sources of more ESs than the remnant forest.•Local people will promote biodiversity conservation if their needs are accommodated.•Agroforests are important as a supplement of forests providing many ESs.
•Mesohabitats on cangas exhibit distinct species richness, abundance, and functional trait patterns, with notable differences between the exposed area and tree association.•Tree association serves as ...a crucial refuge for vulnerable bryophyte species.•A comprehensive and integrated approach is necessary to protect brazilian Cangas, ensuring the survival of plant species in this diverse and heterogeneous environment.
The Canga vegetation in Brazil is a unique ecosystem found on ironstone outcrops, known as ferruginous rupestrian fields, and is considered one of the most diverse in the world. This mosaic landscape, comprised of trees, shrubs, and herbaceous vegetation, is home to a diverse range of plants, including angiosperm, bryophytes, and lichens. To understand the effect of Cangas' patchy landscape, here called mesohabitats, on bryophyte communities, we asked the following questions: Do the diversity parameters such as richness, abundance and composition of bryophytes differ along the different Cangas’ mesohabitats? Do bryophyte assemblages group by functional traits and filter per mesohabitat? We surveyed eight sites in the Quadrilátero Ferrífero (Iron Quadrangle), southeast Brazil. We collected and analysed data on bryophyte diversity and functional traits in three different mesohabitats: exposed areas (EA), shrub associations (SA), and tree associ ations (TA). The diversity of both groups was tested using one-way ANOVA; and functional traits were addressed with a Factorial Analysis of Mixed Data (FAMD). There are significant differences in the diversity of bryophytes among the mesohabitats, with TA having the highest diversity and abundance of liverworts and mosses, followed by SA and EA. The study also found bryophyte assemblies with similar functional traits in similar mesohabitats across the Brazilian Quadrilátero Ferrífero. The research revealed that the different mesohabitats provided important and distinct niches for bryophytes in Cangas, and this threatened ecosystem's high diversity must be considered when developing conservation strategies. This entails land managers adopting effective approaches aimed at improving both the quality and connectivity of habitats, fostering biodiversity, and bolstering the resilience of the ecosystem.
•We evaluated the role of rubber tree plantation on butterfly biodiversity in Brazil.•Plantation harbored a diverse fauna but results depended on management regime.•Management with undergrowth ...suppression led to less diverse butterfly assemblage.•Low herbicide application groves contained more diverse butterfly communities.•Strategies to promote understory growth should be encouraged by industry.
Agroforestry systems have increased in area in tropical regions in recent decades and many studies have sought to evaluate their impact on native biodiversity. Yet, few have assessed the impact of perennial plantations such as rubber-tree harvesting on native biodiversity. The goal of our study was to assess the effect of rubber tree plantations on fruit-feeding butterflies of the endangered Brazilian Atlantic Forest in Brazil. To do so, we sampled fruit- feeding butterfly species in a landscape mosaic composed of primary forest, rubber tree plantations under two management regimes (active production with intense management and undergrowth suppression and low management plantations with no undergrowth suppression), and forest fragments immersed in rubber tree plantation matrix. By trap-baiting butterflies for a year, we captured 5800 individuals of 85 butterfly species. Species richness was higher in unmanaged (no growth suppression) plantation and forest fragments (57–60 species) and lower in managed plantation (with growth suppression) (47) and primary forest (43). Ordination analysis suggests three main community groups formed by primary forest samples, a cluster combining unmanaged plantation and fragments, and managed plantation. There was substantial variation in butterfly abundance in the landscape, but our data suggest that several forest specialists species are able to occur along the mosaic on the landscape, and despite differences in management the entire landscape can contribute for a rich biota. Loss of understory vegetation led to simplified communities, with skewed dominance of a few species. By allowing understory development, a low impact management can provide adequate habitat for native butterflies. Yet, current rubber tree plantation technology does not normally use this method, opting to use the high management approach instead. We hypothesize that these minimally benign plantations may serve as conduits for butterflies in forest patches. Thus, we suggest that rubber tree plantations near Atlantic forest fragments should encourage understory development and establish a landscape mosaic, allowing forest fragments immersed in plantation matrix to be able to exchange individuals and colonize more complex plantation habitat. Intensive suppression of undergrowth should be avoided, except for trail maintenance, and isolation of fragments in this matrix should be kept to a minimum. If undergrowth suppression cannot be avoided, then establishment of stepping stones in plantation matrix should be encouraged to reduce isolation of fragments.
This paper presents the study of modern pollen analogs from the Balearic Islands. While similar studies have been largely applied to mainland areas, research focused on modern vegetation dynamics on ...Mediterranean islands remains very rare. In this research, we combine vegetation surveys, pollen analysis and multivariate statistics to understand landscape composition. The main objectives of are: (1) to examine pollen-vegetation relationships in relation to environmental and land-use variables; (2) to understand modern pollen representation in a mosaic landscape structure; and (3) to propose pollen indicators that characterize the primary vegetation types from the Balearic Islands to better interpret past pollen records in Mediterranean island environments. Pollen results and Redundancy Analysis (RDA) distinguish three major groups: (a) Holm oak and box formations; (b) maquis and garrigues; and (c) anthropogenic and open habitats. Landscape form, mean decadal rainfall, mean decadal temperature, fire activity, trampling, slope percentage, wet/flooded soil, saline soil, distance to agropastoral cells, gHM index, domestic herbivory presence, agropastoral use, and soil type are the major variables explaining modern pollen assemblage variation in our research. Poaceae undiff., Plantago sp., Apiaceae undiff., Cerealia-t, and Cichorieae are highly correlated to human activities but should be interpreted cautiously when occurring in low values. Quercus ilex-t, Hypericum, and Buxus are correlated to humid locations while Pistacia, Pinus, Juniperus-t, and Olea to high mean decadal temperatures. Our study indicates how pollen analysis and multivariate analysis are powerful tools for characterizing the mosaic landscape, with special focus on the main vegetation types of the Balearic Islands.
Boreal forests are globally extensive and store large amounts of carbon, but recent climate change has led to drier conditions and increasing fire activity. The objective of this study is to quantify ...trends in fire size and frequency using data spanning multiple scales in space and time. We use multi-temporal Landsat image compositing on Google Earth Engine and validate results with reference fire maps from the Canadian Park Service. We also interpret general fire trends through the concept of Self-Organized Criticality (SOC). Our study site is Wood Buffalo National Park, which is a fire hot spot in Canada due to frequent lightning ignitions. The relativize differenced normalized burn ratio (RdNBR) was the most accurate Landsat-based burn severity metric we evaluated (52.2% producer's accuracy, 87.6% user's accuracy). The Landsat-based burn severity maps provided a better fit for a linear relationship on the log-log scale of fire size and frequency than a manually drawn fire map. Landsat-based fire trends since 1990 conformed to a power-law distribution with a slope of 1.9, which is related to fractal dimensions of the satellite-based fire perimeter shapes. The unburned and low-severity patches within the burn severity mosaic influenced the power-law slope and associated fractal dimensionality. This study demonstrates a multi-scale and multi-dataset technique to quantify general fire trends and changing fire cycles in remote locations and establishes a baseline database for assessing future fire activity. Testing criticality by power laws helps to quantify emergent trends of contemporary fire regimes, which could inform the strategic application of prescribed fire and other management activities. Natural resource managers can utilize information from this study to understand local ecosystem adaptability to large fire events and ecosystem stability in the context of recent increasing fire activity.
•Image compositing with spectral indices enabled improved mapping of spatiotemporal fire trends.•The multi-temporal fire map created from satellite images followed self-organized criticality.•Fire trends were obtained by a slope in log-log scale of size and frequency of fire.•The slope for fire trends was influenced by fractal dimensions of areas burned by fire.
Dispersal of A. lucorum adult mirid bugs revealed by carbon 13 stable isotope analysis, in Chinese Bt-cotton agro-landscapes, during the growing season of maize and cotton. Important fluxes of mirid ...bugs are measured from C3 host plants (thus, most likely cotton, the biggest reservoir of mirid bugs in the environment) to maize fields (maize in the only C4 plant in the environment) as shown by large, green arrows but reciprocal fluxes are not observed (small dashed, blue arrows). Therefore, maize fields have a potential to act as a sink for outbreaking mirid bug pests from source cotton fields. A, E: vegetative growth of cotton and maize, respectively; B, F: flowering stages of cotton and maize, respectively; C–D, F: fruiting stages of cotton and maize, respectively.
Display omitted
•Tracking the dispersal of arthropod pests is essential to implement informed IPM strategies.•We used carbon 13 stable isotope analysis to estimate the dispersal of adult A. lucorum mirid bugs.•Adults fed on C3- vs. C4-plants as juveniles have different isotopic signatures for up to 9 days after shifting diet.•Most mirid bugs likely fed on cotton as juveniles and moved to maize fields once adults, but reciprocal movements were rare.•Maize fields have a good potential to be a sink for A. lucorum in Chinese Bt-cotton agro-landscapes.
Agricultural landscapes are fast-changing systems due to crop planting and harvesting. These events strongly influence movements of arthropod pests and their natural enemies, yet they are insufficiently considered to implement informed pest management strategies. In China, the adoption of Bt-cotton crops at very large scales has given rise to the recent pest status of mirid bugs such as Apolygus lucorum in cotton fields. In this study we relied on carbon 13 stable isotope analysis to estimate the dispersal of A. lucorum in the Bt-cotton agricultural landscape, composed of a mosaic of cotton and maize and other minor crops. We showed that a diet on C3-plants including cotton induced a significantly different signature than the C4-plant maize. Based on δ13C ratios, we showed that more than 75% of mirid bugs caught in maize fields did not originate in situ, since they had a C3-diet and therefore most likely dispersed from neighbouring cotton fields. Conversely, less than 10% of mirid bugs in cotton fields had a diet on C4-plants, so the vast majority of them did not disperse from maize fields but instead likely stayed in cotton fields. We also observed a delayed establishment of A. lucorum populations in maize fields compared to cotton fields, likely due to host phenology. Indeed we showed in laboratory that juveniles’ survival rate and adults’ fecundity are high on maize silk and grain only, whose resources are available in the field shortly before harvest. Unidirectional movements of mirids from cotton to maize associated with poor developmental success on maize highlight the high potential of maize fields to act as a sink for A. lucorum mirid bug pests. This could be wisely implemented in IPM programs to enhance the management of mirid bug population outbreaks in cotton fields and reduce yield loss on cotton.
Deforestation and forest degradation continue at alarming rates. Landowners often clear forest patches for alternative use, creating landscape mosaics of different land‐use legacies. Ecological ...restoration is usually monitored by aboveground processes, while belowground processes are far less studied. We addressed how the abundance and activity of nutrient improvement bacteria contribute to restoration success, considering different restoration interventions and land‐use legacies. We assessed the abundance and potential activity of N‐fixing and P‐mineralizing bacteria in the rhizosphere of 279 two‐year‐old Quercus mexicana plantlets in a disturbed oak forest in central Mexico. We analyzed the differences across landscape patches (eroded landslide, pastureland, and forest) of four common restoration interventions: passive restoration, artificial shading, topsoil translocation, and combining both treatments. Restoration interventions modified soil functioning differently across landscape patches with a significant increase of P‐mineralizing bacteria abundance and potential activity in the landslide and pastureland. In contrast, N‐fixing bacteria increased only in the landslide site where we registered the lowest organic matter content. Surprisingly, interventions that enhance bacterial activity and abundance do not necessarily improve short‐term seedling performance. We recommend considering the landscape heterogeneity to better match the restoration interventions with the prevailing degradation factor in each landscape patch. Also, combining restoration strategies does not necessarily result in synergy and may imply a useless expense of resources. Finally, monitoring soil health provides novel insights to understand restoration trajectories that remain undetected when focusing only on seedling performance.
PDF
