Diversity–disturbance relationship in forest landscapes

Context

Despite decades of research, there is an intense debate about the consistency of the hump-shaped pattern describing the relationship between diversity and disturbance as predicted by the intermediate disturbance hypothesis (IDH). Previous meta-analyses have not explicitly considered interactive effects of disturbance frequency and intensity of disturbance on plant species diversity in terrestrial landscapes.

Objective

We conducted meta-analyses to test the applicability of IDH by simultaneously examining the relationship between species richness, disturbance frequency (quantified as time since last disturbance as originally proposed) and intensity of disturbance in forest landscapes.

Methods

The effects of disturbance frequency, intensity, and their interaction on species richness was evaluated using a mixed-effects model.

Results

We found that species richness peaks at intermediate frequency after both high and intermediate disturbance intensities, but the richness-frequency relationship differed between intensity classes.

Conclusions

Our study highlights the need to measure multiple disturbance components that could help reconcile conflicting empirical results on the effect of disturbance on plant species diversity.

     

Island biogeography theory outweighs habitat amount hypothesis in predicting plant species richness in small grassland remnants

Context

The habitat amount hypothesis has rarely been tested on plant communities. It remains unclear how habitat amount affect species richness in habitat fragments compared to island effects such as isolation and patch size.

Objectives

How do patch size and spatial distribution compared to habitat amount predict plant species richness and grassland specialist plant species in small grassland remnants? How does sampling area affect the prediction of spatial variables on species richness?

Methods

We recorded plant species density and richness on 131 midfield islets (small remnants of semi-natural grassland) situated in 27 landscapes in Sweden. Further, we tested how habitat amount, compared to focal patch size and distance to nearest neighbor predicted species density and richness of plants and of grassland specialists.

Results

A total of 381 plant species were recorded (including 85 grassland specialist species). A combination of patch size and isolation was better in predicting both density and richness of species compared to habitat amount. Almost 45% of species richness and 23% of specialist species were explained by island biogeography parameters compared to 19 and 11% by the amount of habitat. A scaled sampling method increased the explanation level of island biogeography parameters and habitat amount.

Conclusions

Habitat amount as a concept is not as good as island biogeography to predict species richness in small habitats. Priority in landscape planning should be on larger patches rather than several small, even if they are close together. We recommend a sampling area scaled to patch size in small habitats.

     

Can anthropogenic linear gaps increase plant abundance and diversity?

Context

Seed-dispersing animals often move along “linear gaps” (linear anthropogenic features such as roads and trails that contain little vegetation), especially in densely-vegetated landscapes. As a result, linear gaps and their verges may receive more seeds than adjacent habitats. In addition, linear gap verges may provide more suitable conditions for plant establishment than neighboring habitats. In this way, linear gaps may increase plant abundance and diversity, and facilitate connectivity of native and non-native plant populations, ultimately increasing plant diversity in the landscape.

Objectives

We reviewed current evidence for the potential of anthropogenic linear gaps to increase plant abundance and diversity, and for the mechanisms involved.

Methods

We reviewed peer-reviewed literature published up to December 31st, 2014.

Results

Most (69.2 %) studies found significantly higher plant abundance and/or diversity in linear gap verges than in adjacent habitats. This suggests that linear gaps can increase plant abundance and diversity, and possibly facilitate population spread. However, there was a strong bias toward the study of exotic species. In addition, there were few mechanistic studies to allow estimation of the relative contributions of dispersal and post-dispersal mechanisms operating in linear gaps.

Conclusions

Future studies should focus on entire plant communities, not just exotic species, and should allow identification of the mechanisms by which linear gaps increase plant abundance and diversity. With this knowledge in hand, we will be in a better position to understand whether the net benefit of linear gaps for plant diversity in general outweigh their facilitation of the spread of exotic species.

     

Landscape relatedness: detecting contemporary fine-scale spatial structure in wild populations

Context

Methods for detecting contemporary, fine-scale population genetic structure in continuous populations are scarce. Yet such methods are vital for ecological and conservation studies, particularly under a changing landscape.

Objectives

Here we present a novel, spatially explicit method that we call landscape relatedness (LandRel). With this method, we aim to detect contemporary, fine-scale population structure that is sensitive to spatial and temporal changes in the landscape.

Methods

We interpolate spatially determined relatedness values based on SNP genotypes across the landscape. Interpolations are calculated using the Bayesian inference approach integrated nested Laplace approximation. We empirically tested this method on a continuous population of brown bears (Ursus arctos) spanning two counties in Sweden.

Results

Two areas were identified as differentiated from the remaining population. Further analysis suggests that inbreeding has occurred in at least one of these areas.

Conclusions

LandRel enabled us to identify previously unknown fine-scale structuring in the population. These results will help direct future research efforts, conservation action and aid in the management of the Scandinavian brown bear population. LandRel thus offers an approach for detecting subtle population structure with a focus on contemporary, fine-scale analysis of continuous populations.

     

Distribution of invasive plants in urban environment is strongly spatially structured

Context

Urban environments create a wide range of habitats that harbour a great diversity of plant species, many of which are of alien origin. For future urban planning and management of the green areas within the city, understanding of the spatial distribution of invasive alien species is of great importance.

Objectives

Our main aim was to assess how availability of different ecosystem types within a city area, as well as several parameters describing urban structure interact in determining the cover and identity of invasive alien species.

Methods

We studied the distribution of chosen invasive plant species in a mid-sized city in the Czech Republic, central Europe, on a gradient of equal sized cells from the city centre to its outskirts.

Results

A great amount of variation was explained by spatial predictors but not shared with any measured variables. The species cover of invasive species decreased with increasing proportion of urban greenery and distance from the city centre, but increased with habitat richness; road margins, ruderal sites, and railway sites were richest in invasive species. In contrast, the total number of invasive species in cells significantly decreased with increasing distance from the city centre, but increased with habitat richness.

Conclusions

Our results suggest that different invasive species prefer habitats in the vicinity of the city centre and at its periphery and the spatial structure and habitat quality of the urban landscape needs to be taken into account, in efforts to manage alien plant species invasions in urban environments.

     

Spatial relationship between biodiversity and geodiversity across a gradient of land-use intensity in high-latitude landscapes

Context

‘Conserving Nature’s stage’ has been advanced as an important conservation principle because of known links between biodiversity and abiotic environmental diversity, especially in sensitive high-latitude environments and at the landscape scale. However these links have not been examined across gradients of human impact on the landscape.

Objectives

To (1) analyze the relationships between land-use intensity and both landscape-scale biodiversity and geodiversity, and (2) assess the contributions of geodiversity, climate and spatial variables to explaining vascular plant species richness in landscapes of low, moderate and high human impact.

Methods

We used generalized additive models (GAMs) to analyze relationships between land-use intensity and both geodiversity (geological, geomorphological and hydrological richness) and plant species richness in 6191 1-km² grid squares across Finland. We used linear regression-based variation partitioning (VP) to assess contributions of climate, geodiversity and spatial variable groups to accounting for spatial variation in species richness.

Results

In GAMs, geodiversity correlated negatively, and plant species richness positively, with land-use intensity. Both relationships were non-linear. In VP, geodiversity best accounted for species richness in areas of moderate to high human impact. These overall contributions were mainly due to variation explained jointly with climate, which dominated the models. Independent geodiversity contributions were highest in pristine environments, but low throughout.

Conclusions

Human action increases biodiversity but may reduce geodiversity, at landscape scale in high-latitude environments. Better understanding of the connections between biodiversity and abiotic environment along changing land-use gradients is essential in developing sustainable measures to conserve biodiversity under global change.

     

Landscape genetics of the foundational salt marsh plant species black needlerush (<Emphasis Type="Italic">Juncus roemerianus</Emphasis> Scheele) across the northeastern Gulf of Mexico

Context

Common species important for ecosystem restoration stand to lose as much genetic diversity from anthropogenic habitat fragmentation and climate change as rare species, but are rarely studied. Salt marshes, valuable ecosystems in widespread decline due to human development, are dominated by the foundational plant species black needlerush (Juncus roemerianus Scheele) in the northeastern Gulf of Mexico.

Objectives

We assessed genetic patterns in J. roemerianus by measuring genetic and genotypic diversity, and characterizing population structure. We examined population connectivity by delineating possible dispersal corridors, and identified landscape factors influencing population connectivity.

Methods

A panel of 19 microsatellite markers was used to genotype 576 samples from ten sites across the northeastern Gulf of Mexico from the Grand Bay National Estuarine Research Reserve (NERR) to the Apalachicola NERR. Genetic distances (F_ST and D_ch) were used in a least cost transect analysis (LCTA) within a hierarchical model selection framework.

Results

Genetic and genotypic diversity results were higher than expected based on life history literature, and samples structured into two large, admixed genetic clusters across the study area, indicating sexual reproduction may not be as rare as predicted in this clonal macrophyte. Digitized coastal transects buffered by 500 m may represent possible dispersal corridors, and developed land may significantly impede population connectivity in J. roemerianus.

Conclusions

Results have important implications for coastal restoration and management that seek to preserve adaptive potential by sustaining natural levels of genetic diversity and conserving population connectivity. Our methodology could be applied to other common, widespread and understudied species.

     

Shaken but not stirred: multiscale habitat suitability modeling of sympatric marten species (<Emphasis Type="Italic">Martes martes</Emphasis> and <Emphasis Type="Italic">Martes foina</Emphasis>) in the northern Iberian Peninsula

Context

Multispecies and multiscale habitat suitability models (HSM) are important to identify the environmental variables and scales influencing habitat selection and facilitate the comparison of closely related species with different ecological requirements.

Objectives

This study explores the multiscale relationships of habitat suitability for the pine (Martes martes) and stone marten (M. foina) in northern Spain to evaluate differences in habitat selection and scaling, and to determine if there is habitat niche displacement when both species coexist.

Methods

We combined bivariate scaling and maximum entropy modeling to compare the multiscale habitat selection of the two martens. To optimize the HSM, the performance of three sampling bias correction methods at four spatial scales was explored. HSMs were compared to explore niche differentiation between species through a niche identity test.

Results

The comparison among HSMs resulted in the detection of a significant niche divergence between species. The pine marten was positively associated with cooler mountainous areas, low levels of human disturbance, high proportion of natural forests and well-connected forestry plantations, and medium-extent agroforestry mosaics. The stone marten was positively related to the density of urban areas, the proportion and extensiveness of croplands, the existence of some scrub cover and semi-continuous grasslands.

Conclusions

This study outlines the influence of the spatial scale and the importance of the sampling bias corrections in HSM, and to our knowledge, it is the first comparing multiscale habitat selection and niche divergence of two related marten species. This study provides a useful methodological framework for multispecies and multiscale comparatives.

     

Thresholds and hotspots for shrub restoration following a heterogeneous megafire

Context

Reestablishing foundational plant species through aerial seeding is an essential yet challenging step for restoring the vast semiarid landscapes impacted by plant invasions and wildfire-regime shifts. A key component of the challenge stems from landscape variability and its effects on plant recovery.

Objectives

We assessed landscape correlates, thresholds, and tipping points for sagebrush presence from fine-scale sampling across a large, heterogeneous area burned the previous year, where we were able to quantify soil surface features that are typically occluded yet can strongly affect recovery patterns.

Methods

Hypothesis testing and binary-decision trees were used to evaluate factors affecting initial sagebrush establishment, using 2171 field plots (totaling?~?2,000,000 m² sampled) over a 113,000-ha region.

Results

Sagebrush established in 50% of plots where it was seeded, a?>?12-fold greater establishment frequency than in unseeded areas. Sagebrush establishment was enhanced in threshold-like ways by elevation (>?1200 m ASL), topographic features that alter heatload and soil water, and by soil-surface features such as “fertile islands” that bore the imprint of pre-fire sagebrush. Sagebrush occupancy had a negative, linear relationship with exotic-annual grass cover and parabolic relationship with perennial bunchgrasses (optimal at 40% cover).

Conclusions

Our approach revealed interactive, ecological relationships such as novel soil-surface effects on first year establishment of sagebrush across the burned landscape, and identified “hot spots” for recovery. The approach could be expanded across sites and years to provide the information needed to explain past seeding successes or failures, and in designing treatments at the landscape scale.

     

Landscape structure shapes carnivore-mediated seed dispersal kernels

Context

Seed dispersal is recognized as having profound effects on the distribution, dynamics and structure of plant populations and communities. However, knowledge of how landscape structure shapes carnivore-mediated seed dispersal patterns is still scarce, thereby limiting our understanding of large-scale plant population processes.

Objectives

We aim to determine how the amount and spatial configuration of forest cover impacted the relative abundance of carnivorous mammals, and how these effects cascaded through the seed dispersal kernels they generated.

Methods

Camera traps activated by animal movement were used for carnivore sampling. Colour-coded seed mimics embedded in common figs were used to know the exact origin of the dispersed seed mimics later found in carnivore scats. We applied this procedure in two sites differing in landscape structure.

Results

We did not find between-site differences in the relative abundance of the principal carnivore species contributing to seed dispersal patterns, Martes foina. Mean dispersal distance and the probability of long dispersal events were higher in the site with spatially continuous and abundant forest cover, compared to the site with spatially aggregated and scarcer forest cover. Seed deposition closely matched the spatial patterning of forest cover in both study sites, suggesting behaviour-based mechanisms underpinning seed dispersal patterns generated by individual frugivore species.

Conclusions

Our results provide the first empirical evidence of the impact of landscape structure on carnivore-mediated seed dispersal kernels. They also indicate that seed dispersal kernels generated strongly depend on the effect that landscape structure exerts on carnivore populations, particularly on habitat-use preferences.

     

Butterfly dispersal in farmland: a replicated landscape genetics study on the meadow brown butterfly (<Emphasis Type="Italic">Maniola jurtina</Emphasis>)

Context

Anthropogenic activities readily result in the fragmentation of habitats such that species persistence increasingly depends on their ability to disperse. However, landscape features that enhance or limit individual dispersal are often poorly understood. Landscape genetics has recently provided innovative solutions to evaluate landscape resistance to dispersal.

Objectives

We studied the dispersal of the common meadow brown butterfly, Maniola jurtina, in agricultural landscapes, using a replicated study design and rigorous statistical analyses. Based on existing behavioral and life history research, we hypothesized that the meadow brown would preferentially disperse through its preferred grassy habitats (meadows and road verges) and avoid dispersing through woodlands and the agricultural matrix.

Methods

Samples were collected in 18 study landscapes of 5 × 5 km in three contrasting agricultural French regions. Using circuit theory, least cost path and transect-based methods, we analyzed the effect of the landscape on gene flow separately for each sex.

Results

Analysis of 1681 samples with 6 microsatellites loci revealed that landscape features weakly influence meadow brown butterfly gene flow. Gene flow in both sexes appeared to be weakly limited by forests and arable lands, whereas grasslands and grassy linear elements (road verges) were more likely to enhance gene flow.

Conclusion

Our results are consistent with the hypothesis of greater dispersal through landscape elements that are most similar to suitable habitat. Our spatially replicated landscape genetics study allowed us to detect subtle landscape effects on butterfly gene flow, and these findings were reinforced by consistent results across analytical methods.

     

Deciphering rangeland transformation—complex dynamics obscure interpretations of woody plant encroachment

Context

Woody plant encroachment—the conversion of grasslands to woodlands—is among the greatest challenges faced by rangelands worldwide. Yet this phenomenon is poorly understood, and complex land use dynamics make interpreting the timing and extent of land cover changes a global challenge.

Objectives

For many regions, the true degree and rate of historical change in woody cover and cropland remain unknown. We quantify these processes and explore the effects of prior cultivation on woody plant distribution.

Methods

In the Lampasas Cut Plain, USA, we measured rangeland transformation using digital classification of aerial imagery 1937–2012. Our study is the first to use data of such high spatial and temporal resolution to address this question at this scale. We also provide some of the first documentation of dramatic regional cropland abandonment.

Results

Although total woody cover remained almost unchanged (1937: 28%, 2012: 27 %), woody cover underwent a major redistribution across the landscape. Formerly open areas attained much greater levels of woody cover, and previously wooded areas lost woody cover. As cropland area declined by 78 %, woody plants invaded former croplands more slowly than the rangeland portions of the area (0.1 % year^?1 vs. 0.3 % year^?1, respectively).

Conclusions

These findings conflict with widely held assumptions and suggest that woody plant encroachment is more nuanced than often recognized. Multiple dynamics and past conditions interact in complex ways to produce landscape change. Because perceptions of encroachment determine how we respond to this challenge, great care should be taken in interpreting observed woody plant encroachment of the world’s rangelands.

     

Multi-scale Mexican spotted owl (<Emphasis Type="Italic">Strix occidentalis lucida</Emphasis>) nest/roost habitat selection in Arizona and a comparison with single-scale modeling results

Context

Organisms commonly respond to their environment across a range of scales, however many habitat selection studies still conduct selection analyses using a single-scale framework. The adoption of multi-scale modeling frameworks in habitat selection studies can improve the effectiveness of these studies and provide greater insights into scale-dependent relationships between species and specific habitat components.

Objectives

Our study assessed multi-scale nest/roost habitat selection of the federally “Threatened” Mexican spotted owl (Strix occidentalis lucida) in northern Arizona, USA in an effort to provide improved conservation and management strategies for this subspecies.

Methods

We conducted multi-scale habitat modeling to assess habitat selection by Mexican spotted owls using survey data collected by the USFS. Each selected covariate was included in multi-scale models at their “characteristic scale” and we used an all-subsets approach and model selection framework to assess habitat selection.

Results

The “characteristic scale” identified for each covariate varied considerably among covariates and results from multi-scale models indicated that percent canopy cover and slope were the most important covariates with respect to habitat selection by Mexican spotted owls. Multi-scale models consistently outperformed their analogous single-scale counterparts with respect to the proportion of deviance explained and model predictive performance.

Conclusions

Efficacy of future habitat selection studies will benefit by taking a multi-scale approach. In addition to potentially providing increased explanatory power and predictive capacity, multi-scale habitat models enhance our understanding of the scales at which species respond to their environment, which is critical knowledge required to implement effective conservation and management strategies.

     

Synthetic Aperture Radar (SAR) images improve habitat suitability models

Context

The ability to detect ecological networks in landscapes is of utmost importance for managing biodiversity and planning corridors.

Objectives

The objective of this study was to evaluate the information provided by a synthetic aperture radar (SAR) image for landscape connectivity modeling compared to aerial photographs (APs).

Methods

We present a novel method that integrates habitat suitability derived from remote sensing imagery into a connectivity model to explain species abundance. More precisely, we compared how two resistance maps constructed using landscape and/or local metrics derived from AP or SAR imagery yield different connectivity values (based on graph theory), considering hedgerow networks and forest carabid beetle species as a model.

Results

We found that resistance maps using landscape and local metrics derived from SAR imagery improve landscape connectivity measures. The SAR model is the most informative, explaining 58% of the variance in forest carabid beetle abundance. This model calculates resistance values associated with homogeneous patches within hedgerows according to their suitability (canopy cover density and landscape grain) for the model species.

Conclusions

Our approach combines two important methods in landscape ecology: the construction of resistance maps and the use of buffers around sampling points to determine the importance of landscape factors. This study was carried out through an interdisciplinary approach involving remote sensing scientists and landscape ecologists. This study is a step forward in developing landscape metrics from satellites to monitor biodiversity.

     

The plant functional traits that explain species occurrence across fragmented grasslands differ according to patch management,isolation, and wetness

Context

Landscape fragmentation significantly affects species distributions by decreasing the number and connectivity of suitable patches. While researchers have hypothesized that species functional traits could help in predicting species distribution in a landscape, predictions should depend on the type of patches available and on the ability of species to disperse and grow there.

Objectives

To explore whether different traits can explain the frequency of grassland species (number of occupied patches) and/or their occupancy (ratio of occupied to suitable patches) across a variety of patch types within a fragmented landscape.

Methods

We sampled species distributions over 1300 grassland patches in a fragmented landscape of 385 km² in the Czech Republic. Relationships between functional traits and species frequency and occupancy were tested across all patches in the landscape, as well as within patches that shared similar management, wetness, and isolation.

Results

Although some traits predicting species frequency also predicted occupancy, others were markedly different, with competition- and dispersal-related traits becoming more important for occupancy. Which traits were important differed for frequency and occupancy and also differed depending on patch management, wetness, and isolation.

Conclusions

Plant traits can provide insight into plant distribution in fragmented landscapes and can reveal specific abiotic, biotic, and dispersal processes affecting species occurrence in a patch type. However, the importance of individual traits depends on the type of suitable patches available within the landscape.

     

Landscape composition mediates movement and habitat selection in bobcats (<Emphasis Type="Italic">Lynx rufus</Emphasis>): implications for conservation planning

Context

The analysis of individual movement choices can be used to better understand population-level resource selection and inform management.

Objectives

We investigated movements and habitat selection of 13 bobcats in Vermont, USA, under the assumption individuals makes choices based upon their current location. Results were used to identify “movement-defined” corridors.

Methods

We used GPS-collars and GIS to estimate bobcat movement paths, and extracted statistics on land cover proportions, topography, fine-scale vegetation, roads, and streams within “used” and “available” space surrounding each movement path. Compositional analyses were used to determine habitat preferences with respect to landcover and topography; ratio tests were used to determine if used versus available ratios for vegetation, roads, and streams differed from 1. Results were used to create travel cost maps, a primary input for corridor analysis.

Results

Forested and scrub-rock land cover were most preferred for movement, while developed land cover was least preferred. Preference depended on the composition of the “available” landscape: Bobcats moved?>?3 times more quickly through forest and scrub-rock habitat when these habitats were surrounded by agriculture or development than when the available buffer was similarly composed. Overall, forest edge, wetland edge and higher stream densities were selected, while deep forest core and high road densities were not selected. Landscape-scale connectivity maps differed depending on whether habitat suitability, preference, or selection informed the travel cost map.

Conclusions

Both local and landscape scale land cover characteristics affect habitat preferences and travel speed of bobcats, which in turn can inform management and conservation activities.

     

Contemporary spatial and environmental factors determine vascular plant species richness on highly fragmented meadows in Central Finland

Context

Habitat loss is a major threat to biodiversity. It can create temporal lags in decline of species in relation to destruction of habitat coverage. Plant species specialized in semi-natural grasslands, especially meadows, often express such extinction debt.

Objectives

We studied habitat loss and fragmentation of meadows and examined whether the changes in meadow coverage had caused an extinction debt on vascular plants. We also studied whether historical or present landscape patterns or contemporary environmental factors were more important determinants of species occurrence.

Methods

We surveyed the plant species assemblages of 12 grazed and 12 mown meadows in Central Finland and detected the meadow coverages from their surroundings on two spatial scales and on three time steps. We modelled the effects of functional connectivity, habitat amount, and isolation on species richness and community composition.

Results

We observed drastic and dynamic meadow loss in landscapes surrounding our study sites during the last 150 years. However, we did not find explicit evidence for an extinction debt in meadow plants. The observed species richness correlated with contemporary factors, whereas both contemporary factors and habitat availability during the 1960s affected community composition.

Conclusions

Effective conservation management of meadow biodiversity builds on accurate understanding of the relative importance of past and present factors on species assemblages. Both mown and grazed meadows with high species richness need to be managed in the future. The management effort should preferably be targeted to sites located near to each other.

     

Spatial assessment of aesthetic services in a complex mountain region: combining visual landscape properties with crowdsourced geographic information

Context

Human and natural systems interact at multiple scales which are context specific in relation to ecosystem service supply. Scenic beauty is recognised as a cultural ecosystem service whose aesthetic value is perceived at a holistic landscape level.

Objectives

In this study we provide methodological advancements for assessing the relationship between landscape visual character and scenic beauty based on crowdsourced geographic information. The final aim is to demonstrate, through a case study application, an empirical method for mapping the scenic beauty of complex mountain landscapes from the perspective of observers which are realistically exposed to the environment being evaluated.

Methods

We propose a viewshed based approach which relies on visual indicators and the location of visitors retrieved by public image storage analysis. A cluster analysis was used to integrate visual characters of the landscape and visiting users’ preferences.

Results

Four different typologies of landscapes were finally characterized by distinct values of visual indicators. The spatial distribution of the landscape typologies presented a clustered pattern, allowing a regionalization of the landscape characters. The analysis of the visiting users’ provenance revealed that visual scale, naturalness and ephemera attract mainly foreign users, while imageability, complexity and historicity attract mostly domestic and local users.

Conclusions

The combination of crowdsourced images with visual indicators allows a systematic analysis of landscape scenic beauty properties. In all, by understanding how specific landscape characters contributes to aesthetic service provision we provide a tool for facilitating the visualization and interpretation of complex landscape characters.

     

Multi-feature machine learning model for automatic segmentation of green fractional vegetation cover for high-throughput field phenotyping

Background

Accurately segmenting vegetation from the background within digital images is both a fundamental and a challenging task in phenotyping. The performance of traditional methods is satisfactory in homogeneous environments, however, performance decreases when applied to images acquired in dynamic field environments.

Results

In this paper, a multi-feature learning method is proposed to quantify vegetation growth in outdoor field conditions. The introduced technique is compared with the state-of the-art and other learning methods on digital images. All methods are compared and evaluated with different environmental conditions and the following criteria: (1) comparison with ground-truth images, (2) variation along a day with changes in ambient illumination, (3) comparison with manual measurements and (4) an estimation of performance along the full life cycle of a wheat canopy.

Conclusion

The method described is capable of coping with the environmental challenges faced in field conditions, with high levels of adaptiveness and without the need for adjusting a threshold for each digital image. The proposed method is also an ideal candidate to process a time series of phenotypic information throughout the crop growth acquired in the field. Moreover, the introduced method has an advantage that it is not limited to growth measurements only but can be applied on other applications such as identifying weeds, diseases, stress, etc.