Search

Accurate and comparable annual mapping is critical to understanding changing vegetation distribution and informing land use planning and management. A U-Net convolutional neural network (CNN) model was used to map natural vegetation and forest types based on annual Landsat geomedian reflectance composite images for a 500 km × 500 km study area in southeastern Australia. The CNN was developed using 2018 imagery. Label data were a ten-class natural vegetation and forest classification (i.e., Acacia, Callitris, Casuarina, Eucalyptus, Grassland, Mangrove, Melaleuca, Plantation, Rainforest and Non-Forest) derived by combining current best-available regional-scale maps of Australian forest types, natural vegetation and land use. The best CNN generated using six Landsat geomedian bands as input produced better results than a pixel-based random forest algorithm, with higher overall accuracy (OA) and weighted mean F1 score for all vegetation classes (93 vs. 87% in both cases) and a higher Kappa score (86 vs. 74%). The trained CNN was used to generate annual vegetation maps for 2000–2019 and evaluated for an independent test area of 100 km × 100 km using statistics describing accuracy regarding the label data and temporal stability. Seventy-six percent of pixels did not change over the 20 years (2000–2019), and year-on-year results were highly correlated (94–97% OA). The accuracy of the CNN model was further verified for the study area using 3456 independent vegetation survey plots where the species of interest had ≥ 50% crown cover. The CNN showed an 81% OA compared with the plot data. The model accuracy was also higher than the label data (76%), which suggests that imperfect training data may not be a major obstacle to CNN-based mapping. Applying the CNN to other regions would help to test the spatial transferability of these techniques and whether they can support the automated production of accurate and comparable annual maps of natural vegetation and forest types required for national reporting. [ABSTRACT FROM AUTHOR]

Land cover mapping from satellite images has progressed from visual and statistical approaches to Random Forests (RFs) and, more recently, advanced image recognition techniques such as convolutional neural networks (CNNs). CNNs have a conceptual benefit over RFs in recognising spatial feature context, but potentially at the cost of reduced spatial detail. We tested the use of CNNs for improved land cover mapping based on Landsat data, compared with RFs, for a study area of approximately 500 km × 500 km in southeastern Australia. Landsat 8 geomedian composite surface reflectances were available for 2018. Label data were a simple nine-member land cover classification derived from reference land use mapping (Catchment Scale Land Use of Australia—CLUM), and further enhanced by using custom forest extent mapping (Forests of Australia). Experiments were undertaken testing U-Net CNN for segmentation of Landsat 8 geomedian imagery to determine the optimal combination of input Landsat 8 bands. The results were compared with those from a simple autoencoder as well as an RF model. Segmentation test results for the best performing U-Net CNN models produced an overall accuracy of 79% and weighted-mean F1 score of 77% (9 band input) or 76% (6 band input) for a simple nine-member land cover classification, compared with 73% and 68% (6 band input), respectively, for the best RF model. We conclude that U-Net CNN models can generate annual land cover maps with good accuracy from proxy training data, and can also be used for quality control or improvement of existing land cover products. [ABSTRACT FROM AUTHOR]

A basalt lava plain covers much of central and western Victoria in Australia. Many of the younger flows (<1 Ma) have hummocky surfaces with abundant tumuli and lava rises, known locally as stony rises due to the protruding basalt boulders. Stony rises are noteworthy geologically but have ecological importance and are culturally significant to Aboriginal Australians. Ecologically, stony rises host many native plant species. Culturally, stony rises provide evidence of Aboriginal land use and occupation through the presence of artefact deposits. Currently, mapping of stony rises is carried out by conducting field surveys on a study-by-study basis. This could be performed more rapidly using remote sensing datasets such as LiDAR, satellite and aerial imagery, and geophysical data. This research mapped stony rises in an area that is experiencing rapid urban development using a combination of remotely sensed datasets that characterise the morphology of stony rises through a Machine Learning approach that utilises Object Based Image Analysis (OBIA) segmentation and Random Forests (RF) classification techniques. The research successfully identified stony rises in the landscape to an accuracy of 89.97 %, with 314 potential new stony rise sites identified. The research identified the most important predictor variables were slope, local elevation, the concentrations of thorium and uranium (obtained from airborne radiometric data), aeromagnetic data (total magnetic intensity), the Normalised Difference Water Index (NDWI), and the Clay Mineral Ratio. The results from this study lay a foundation for standardising the definition of a stony rise within Australia. Furthermore, by utilising various predictor variables, this study highlights the success of a multi-faceted approach into detecting a specific landform at a landscape level. • Stony rises detected to an accuracy of 89.97 % using our novel method. • 314 potential new stony rises were detected using our novel method. • Using multifaceted predictor variables for landform classification and detection [ABSTRACT FROM AUTHOR]

To understand how ecological communities will respond to global change we need new tools and datasets on species across large spatial and temporal scales. Hyperspectral reflectance 'spectra' capture a promising set of traits that show potential to be scaled up in time and space via remote sensing. Thus far, spectra have been shown to distinguish the taxa and trait responses of a substantial number of species within a plethora of vascular plant communities, but not yet for biological soil crust communities (biocrusts).Here, we assess if spectra can be applied to identify biocrust species and their trait variation. We collected biocrust specimens across an aridity gradient spanning 650 km within drylands of Eastern Australia and acquired their spectra, over 12,700 spectral readings, with a high‐resolution radiospectrometer. A machine learning method (random forests) was used to assess how well the spectra of biocrust specimens could distinguish their species and broader structural and chemical traits.Spectra were able to differentiate a substantial number of biocrust species (35) with considerable accuracy (~78.5%). Furthermore, spectral features related to chemical traits were found to primarily drive species spectral differences.Synthesis. Our findings establish that biocrust species hold unique and detectable spectral responses, providing an important basis for remote sensing applications on biocrust species and their trait responses across dryland systems. [ABSTRACT FROM AUTHOR]

Summary: Serrated Tussock (Nassella trichotoma) is a weed that invades native grasslands in southern Australia. It is frequently controlled using extensive and costly chemical control. We describe a tool aimed at grassland reserve managers that provides guidance about the suitability of different chemical control techniques, including knapsack, boom and aerial spraying. The tool takes the attributes of any site and predicts the suitability of each technique. The tool was built using random forests trained on a dataset of practitioner judgements. We discuss the applications and limitations of the tool. [ABSTRACT FROM AUTHOR]

Forest fire severity influences post-fire fuel structure and thus the behaviour of subsequent fires. Understanding such interactions is critical to improving predictions of fire risk and emergency management, yet few studies have quantified fire severity effects on fuel attributes. We quantify fuel structure of a fire-tolerant eucalypt forest 7 years after a landscape-scale wildfire in south-eastern Australia. We used high-density airborne lidar data to estimate understorey fuel metrics in three strata representing horizontal and vertical connectivity in 1084 plots (0.06 ha) representing four wildfire severities (unburnt, low, moderate, high). Fuel structure was changed by high-severity fire, which significantly increased the cover and horizontal connectivity of the elevated and midstorey strata and decreased space between the understorey and canopy relative to other severity types. Random Forest models indicated that understorey fuel metrics were most influenced by wildfire severity, pre-fire values of each metric, and post-fire canopy cover, and least influenced by climatic and topographic variables. Our study provides evidence of positive feedbacks to flammability by high-severity wildfire in fire-tolerant eucalypt forests through increased horizontal and vertical fuel connectivity. It demonstrates the utility of airborne lidar data for quantifying fuel structure in complex forests and providing critical data for fire risk assessments. We use airborne lidar data to examine how wildfire severity influences the horizontal and vertical connectivity of fuel in temperate eucalypt forests. Our study provides evidence of positive feedbacks to flammability and a greater potential for crown fires through increased ladder fuels after high-severity fire. [ABSTRACT FROM AUTHOR]

The source of sedimentary organic matter in lakes can help to elucidate climate and catchment variation and processes that reflect lake development. Common techniques for tracing sediment organic matter sources, such as the stable isotopes and elemental concentrations of C and N, can be too imprecise to identify the specific provenance of organic matter. By contrast, organic geochemical techniques such as gas or liquid chromatography and nuclear magnetic resonance provide detailed organic molecular characterisation but are both expensive and time consuming. Fourier Transform Infrared (FTIR) spectroscopy is a rapid, non-destructive, and well-established method for determining the constituents of lake sediments. However, the potential for identifying the sources of organic matter in lake sediments has not been fully explored. In this study, we assess the extent to which FTIR can be used to identify varying organic matter sources through analysis of modern autotrophs from Blue Lake, North Stradbroke Island, Australia. We investigated spectral processing techniques to identify the approach that could most accurately classify autotroph samples. Three autotroph groups (i.e algae, aquatic macrophytes, and terrestrial) were correctly classified 90% of the time. Processed spectra then became the basis of a model that used multivariate random forests to estimate sediment organic matter composition source from a sediment record from Blue Lake that spans the last 7500 years. FTIR-based estimates suggested that throughout the history of the lake, algae contributed the highest amount of organic matter to the sediment samples. These results allow a refinement of a previous study of C:N and δ13C from the same core and suggests that alterations in C:N and, particularly, δ13C reflect chemical changes in algae through time. This study demonstrates that FTIR spectroscopy is a promising tool to elucidate sources of sediment organic matter in lake sediments. • First attempt to predict sediment organic matter origin using FTIR spectroscopy. • Classification success of autotroph samples was 90%. • A random forest model was used to estimate sediment organic matter origin. • Model estimates more accurate with one source dominant, less accurate with multiple sources. • Sediment estimates indicate algal dominance throughout Blue Lake history. [ABSTRACT FROM AUTHOR]