Explore the vast range of titles available.

Convolutional neural networks (CNN) for spectroscopic modelling are currently tuned manually, and the effects of their hyperparameters are not analysed. These can result in sub-optimal models. Here, we propose an approach to tune one-dimensional CNN (1D-CNNs) automatically. It consists of a parametric representation of 1D-CNNs and an optimisation of hyperparameters to maximise a model’s performance. We used a large European soil spectroscopic database to demonstrate our approach for estimating soil organic carbon (SOC) contents. To assess the optimisation, we compared it to random search, and to understand the effects of the hyperparameters, we calculated their importance using functional Analysis of Variance. Compared to random search, the optimisation produced better final results and showed faster convergence. The optimal model produced the most accurate estimates of SOC with RMSE = 9.67 ± 0.51 (s.d.) and R 2 = 0.89 ± 0.013 (s.d.). The hyperparameters associated with model training and architecture critically affected the model’s performance, while those related to the spectral preprocessing had little effect. The optimisation searched through a complex hyperparameter space and returned an optimal 1D-CNN. Our approach simplified the development of 1D-CNNs for spectroscopic modelling by automatically selecting hyperparameters and preprocessing methods. Hyperparameter importance analysis shed light on the tuning process and increased the model’s reliability.

Soil organic carbon (C) is an essential component of the global C cycle. Processes that control its composition and dynamics over large scales are not well understood. Thus, our understanding of C cycling is incomplete, which makes it difficult to predict C gains and losses due to changes in climate, land use and management. Here we show that controls on the composition of organic C, the particulate, humus and resistant fractions, and the potential vulnerability of C to decomposition across Australia are distinct, scale-dependent and variable. We used machine-learning with 5,721 topsoil measurements to show that, continentally, the climate, soil properties (for example, total nitrogen and pH) and elevation are dominant controls. However, we found that such general assessments disregard underlying region-specific controls that affect the distribution of the organic C fractions and vulnerability. This can lead to misinterpretations that prejudice our understanding of soil C processes and dynamics. Regionally, climate is mediated through interactions with soil properties, mineralogy and topography. In some regions, climate is uninfluential. These results highlight the need for regional assessments of soil C dynamics and more local parameterization of biogeochemical and Earth system models. Our analysis propounds the development of region-specific strategies for effective C management and climate change mitigation.Soil geochemistry can be more important than climate in controlling carbon storage, its composition as well as stability, but controls are distinct, scale-dependent and variable, according to an analysis of topsoil measurements across Australia.

Soil properties, such as organic carbon, pH and clay content, are critical indicators of ecosystem function. Visible–near infrared (vis–NIR) reflectance spectroscopy has been widely used to cost-efficiently estimate such soil properties. Multivariate modelling, such as partial least squares regression (PLSR), and machine learning are the most common methods for modelling soil properties with spectra. Often, such models do not account for the multiresolution information presented in the vis–NIR signal, or the spatial variation in the data. To address these potential shortcomings, we used wavelets to decompose the vis–NIR spectra of 226 soils from agricultural and forested regions in south-western Western Australia and developed a wavelet geographically weighted regression (WGWR) for estimating soil organic carbon content, clay content and pH. To evaluate the WGWR models, we compared them to linear models derived with multiresolution data from a wavelet decomposition (WLR) and PLSR without multiresolution information. Overall, validation of the WGWR models produced more accurate estimates of the soil properties than WLR and PLSR. Around 3.5–49.1% of the improvement in the estimates was due to the multiresolution analysis and 1.0–5.2% due to the integration of spatial information in the modelling. The WGWR improves the modelling of soil properties with spectra.

Clay minerals are the most reactive inorganic components of soils. They help to determine soil properties and largely govern their behaviors and functions. Clay minerals also play important roles in biogeochemical cycling and interact with the environment to affect geomorphic processes such as weathering, erosion and deposition. This paper provides new spatially explicit clay mineralogy information for Australia that will help to improve our understanding of soils and their role in the functioning of landscapes and ecosystems. I measured the abundances of kaolinite, illite and smectite in Australian soils using near infrared (NIR) spectroscopy. Using a model‐tree algorithm, I built rule‐based models for each mineral at two depths (0–20 cm, 60–80 cm) as a function of predictors that represent the soil‐forming factors (climate, parent material, relief, vegetation and time), their processes and the scales at which they vary. The results show that climate, parent material and soil type exert the largest influence on the abundance and spatial distribution of the clay minerals; relief and vegetation have more local effects. I digitally mapped each mineral on a 3 arc‐second grid. The maps show the relative abundances and distributions of kaolinite, illite and smectite in Australian soils. Kaolinite occurs in a range of climates but dominates in deeply weathered soils, in soils of higher landscapes and in regions with more rain. Illite is present in varied landscapes and may be representative of colder, more arid climates, but may also be present in warmer and wetter soil environments. Smectite is often an authigenic mineral, formed from the weathering of basalt, but it also occurs on sediments and calcareous substrates. It occurs predominantly in drier climates and in landscapes with low relief. These new clay mineral maps fill a significant gap in the availability of soil mineralogical information. They provide data to for example, assist with research into soil fertility and food production, carbon sequestration, land degradation, dust and climate modeling and paleoclimatic change. Key Points New insights into factors affecting formation and distribution of clay minerals New methodology for measuring and high‐resolution mapping of clay minerals Can be used by various geoscientists for dust modeling, C seq., food production

Information about the carbon cycle potentially constrains the water cycle, and vice versa. This paper explores the utility of multiple observation sets to constrain a land surface model of Australian terrestrial carbon and water cycles, and the resulting mean carbon pools and fluxes, as well as their temporal and spatial variability. Observations include streamflow from 416 gauged catchments, measurements of evapotranspiration (ET) and net ecosystem production (NEP) from 12 eddy-flux sites, litterfall data, and data on carbon pools. By projecting residuals between observations and corresponding predictions onto uncertainty in model predictions at the continental scale, we find that eddy flux measurements provide a significantly tighter constraint on continental net primary production (NPP) than the other data types. Nonetheless, simultaneous constraint by multiple data types is important for mitigating bias from any single type. Four significant results emerging from the multiply-constrained model are that, for the 1990–2011 period: (i) on the Australian continent, a predominantly semi-arid region, over half the water loss through ET (0.64 ± 0.05) occurs through soil evaporation and bypasses plants entirely; (ii) mean Australian NPP is quantified at 2.2 ± 0.4 (1σ) Pg C yr−1; (iii) annually cyclic (\"grassy\") vegetation and persistent (\"woody\") vegetation account for 0.67 ± 0.14 and 0.33 ± 0.14, respectively, of NPP across Australia; (iv) the average interannual variability of Australia's NEP (±0.18 Pg C yr−1, 1σ) is larger than Australia's total anthropogenic greenhouse gas emissions in 2011 (0.149 Pg C equivalent yr–1), and is dominated by variability in desert and savanna regions.

Our objective was to develop a methodology to predict soil fertility using visible–near-infrared (vis–NIR) diffuse reflectance spectra and terrain attributes derived from a digital elevation model (DEM). Specifically, our aims were to: (i) assemble a minimum data set to develop a soil fertility index for sugarcane (Saccharum officinarum L.) (SFI-SC) for biofuel production in tropical soils; (ii) construct a model to predict the SFI-SC using soil vis–NIR spectra and terrain attributes; and (iii) produce a soil fertility map for our study area and assess it by comparing it with a green vegetation index (GVI). The study area was 185 ha located in São Paulo State, Brazil. In total, 184 soil samples were collected and analyzed for a range of soil chemical and physical properties. Their vis–NIR spectra were collected from 400 to 2500 nm. The Shuttle Radar Topographic Mission 3-arcsec (90-m resolution) DEM of the area was used to derive 17 terrain attributes. A minimum data set of soil properties was selected to develop the SFI-SC. The SFI-SC consisted of three classes: Class 1, the highly fertile soils; Class 2, the fertile soils; and Class 3, the least fertile soils. It was derived heuristically with conditionals and using expert knowledge. The index was modeled with the spectra and terrain data using cross-validated decision trees. The cross-validation of the model correctly predicted Class 1 in 75% of cases, Class 2 in 61%, and Class 3 in 65%. A fertility map was derived for the study area and compared with a map of the GVI. Our approach offers a methodology that incorporates expert knowledge to derive the SFI-SC and uses a versatile spectro-spatial methodology that may be implemented for rapid and accurate determination of soil fertility and better exploration of areas suitable for production.

Understanding the change in soil organic carbon (C) stock in a warmer climate and the effect of current land management on that stock is critical for soil and environmental conservation and climate policy. By simulation modeling, we predicted changes in Australia’s soil organic C stock from 2010 to 2100. These vary from losses of 0.014–0.077 t C ha −1 year −1 between 2020 and 2045 and 0.013–0.047 t C ha −1 year −1 between 2070 and 2100, under increasing emissions of greenhouse gases and temperature. Thus, Australian soil will be a net emitter of CO 2 . Depending on the future socio-economic conditions, we predict that croplands will accrue as much as 0.19 t C ha −1 year −1 between 2020 and 2045 due to their management, but accrual will decrease with warming and increased emissions by 2070–2100. The gains will be too small to counteract the losses of C from the larger areas of rangelands and coastal regions that are more sensitive to a warmer climate. In principle, prudent management of the rangelands, for example, improving grazing management and regenerating biodiverse, endemic native plant communities, could sequester more C and mitigate the loss; in practice, it may be more difficult, requiring innovation, interdisciplinary science, cultural awareness and effective policies.

The debate remains unresolved about soil erosion substantially offsetting fossil fuel emissions and acting as an important source or sink of CO2. There is little historical land use and management context to this debate, which is central to Australia's recent past of European settlement, agricultural expansion and agriculturally induced soil erosion. We use \"catchmen\" scale (25 km2) estimates of 137Cs-derived net (1950s–1990) soil redistribution of all processes (wind, water and tillage) to calculate the net soil organic carbon (SOC) redistribution across Australia. We approximate the selective removal of SOC at net eroding locations and SOC enrichment of transported sediment and net depositional locations. We map net (1950s–1990) SOC redistribution across Australia and estimate erosion by all processes to be 4 Tg SOC yr-1, which represents a loss of 2% of the total carbon stock (0–10 cm) of Australia. Assuming this net SOC loss is mineralised, the flux (15 TgCO2-equivalents yr-1) represents an omitted 12% of CO2-equivalent emissions from all carbon pools in Australia. Although a small source of uncertainty in the Australian carbon budget, the mass flux interacts with energy and water fluxes, and its omission from land surface models likely creates more uncertainty than has been previously recognized.

Iron (Fe) oxide mineralogy in most Australian soils is poorly characterized, even though Fe oxides play an important role in soil function. Fe oxides reflect the conditions of pH, redox potential, moisture, and temperature in the soil environment. The strong pigmenting effect of Fe oxides gives most soils their color, which is largely a reflection of the soil's Fe mineralogy. Visible–near‐infrared (vis–NIR) spectroscopy can be used to identify and measure the abundance of certain Fe oxides in soil, and the visible range can be used to derive tristimuli soil color information. The aims of this paper are (1) to measure the abundance of hematite and goethite in Australian soils from their vis–NIR spectra, (2) to compare these results to measurements of soil color, and (3) to describe the spatial variability of hematite, goethite, and soil color and map their distribution across Australia. We measured the spectra of 4606 surface soil samples from across Australia using a vis–NIR spectrometer with a wavelength range of 350–2500 nm. We determined the Fe oxide abundance for each sample using the diagnostic absorption features of hematite (near 880 nm) and goethite (near 920 nm) and derived a normalized iron oxide difference index (NIODI) to better discriminate between them. The NIODI was generalized across Australia with its spatial uncertainty using sequential indicator simulation, which resulted in a map of the probability of the occurrence of hematite and goethite. We also derived soil RGB color from the spectra and mapped its distribution and uncertainty across the country using sequential Gaussian simulations. The simulated RGB color values were made into a composite true color image and were also converted to Munsell hue, value, and chroma. These color maps were compared to the map of the NIODI, and both were used to interpret our results. The work presented here was validated by randomly splitting the data into training and test data sets, as well as by comparing our results to existing studies on the distribution of Fe oxides in Australian soils.

In rare cases, the removal of a single proton (Z) or neutron (N) from an atomic nucleus leads to a dramatic shape change. These instances are crucial for understanding the components of the nuclear interactions that drive deformation. The mercury isotopes (Z = 80) are a striking example1,2: their close neighbours, the lead isotopes (Z = 82), are spherical and steadily shrink with decreasing N. The even-mass (A = N + Z) mercury isotopes follow this trend. The odd-mass mercury isotopes 181,183,185Hg, however, exhibit noticeably larger charge radii. Due to the experimental difficulties of probing extremely neutron-deficient systems, and the computational complexity of modelling such heavy nuclides, the microscopic origin of this unique shape staggering has remained unclear. Here, by applying resonance ionization spectroscopy, mass spectrometry and nuclear spectroscopy as far as 177Hg, we determine 181Hg as the shape-staggering endpoint. By combining our experimental measurements with Monte Carlo shell model calculations, we conclude that this phenomenon results from the interplay between monopole and quadrupole interactions driving a quantum phase transition, for which we identify the participating orbitals. Although shape staggering in the mercury isotopes is a unique and localized feature in the nuclear chart, it nicely illustrates the concurrence of single-particle and collective degrees of freedom at play in atomic nuclei.