ALBERT

Description: Production and reduction of nitrous oxide (N2O) by soil denitrifiers influences atmospheric concentrations of this potent greenhouse gas. Accurate climate projections of net N2O flux have three key uncertainties: (1) short- vs. long-term responses to warming; (2) interactions among soil horizons; and (3) temperature responses of different steps in the denitrification pathway. We addressed these uncertainties by sampling soil from a boreal forest climate transect encompassing a 5.2 °C difference in mean annual temperature, and incubating the soil horizons in isolation and together at three ecologically relevant temperatures in conditions that promote denitrification. Both short-term exposure to warmer temperatures and long-term exposure to a warmer climate increased N2O emissions from organic and mineral soils; an isotopic tracer suggested an increase in N2O production was more important than a decline in N2O reduction. Short-term warming promoted reduction of organic horizon-derived N2O by mineral soil when these horizons were incubated together. The abundance of nirS (a precursor gene for N2O production) was not sensitive to temperature, while that of nosZ clade I (a gene for N2O reduction) decreased with short-term warming in both horizons and was higher from a warmer climate. These results suggest a decoupling of gene abundance and process rates in these soils that differs across horizons and timescales. In spite of these variations, our results suggest a consistent, positive response of denitrifier-mediated, net N2O efflux rates to temperature across timescales in these boreal forests. Our work also highlights the importance of understanding cross-horizon N2O fluxes for developing a predictive understanding of net N2O efflux from soils.

Description: The number of samples used in the calibration dataset affects the quality of the generated predictive models using visible, near and shortwave infrared (VIS-NIR-SWIR) spectroscopy for soil attributes. Recently, convolutional neural network (CNN) is regarded as a highly accurate model for predicting soil properties on a large database, however it has not been ascertained yet how large the sample size should be for CNN model to be effective. This paper aims at providing an estimate of how much calibration samples are needed to improve the model performance of soil properties predictions with CNN. It is hypothesized that the larger the amount of data, the more accurate is the CNN model. The performance of two commonly used machine learning models (Partial least squares regression (PLSR) and Cubist) are compared against the CNN model. A VIS-NIR-SWIR spectral library from Brazil containing 4251 unique sites, with averages of 2–3 samples per depth (a total of 12,044 samples), was divided into calibration (3188 sites) and validation (1063 sites) sets. A subset of the calibration dataset was then created to represent smaller calibration dataset ranging from 125, 300, 500, 1000, 1500, 2000, 2500 and 2700 unique sites, or equivalent to sample size approximately 350, 840, 1400, 2800, 4200, 5600, 7000, and 7650. All three models (PLSR, Cubist, and CNN models) were generated for each sample size of the unique sites for the prediction of five different soil properties, i.e. cation exchange capacity, organic matter, sand, silt and clay content. These calibration subset sampling processes and modelling were repeated ten times to provide a better representation of the model performances. Similar results were observed when the performances of both PLSR and Cubist model were compared to the CNN model where the performance of CNN outweighed the PLSR and Cubist model at sample size of 1500 and 1800 respectively. It can be recommended that deep learning is most efficient for spectral modelling for sample size above 2000. The accuracy of the PLSR and Cubist model seemed to reach a plateau above sample size of 4200 and 5000 respectively. A sensitivity analysis was performed on the CNN model to determine important wavelengths region that affected the predictions of various soil attributes.

Description: The application of machine learning (ML) techniques in various fields of science has increased rapidly, especially in the last ten years. The increasing availability of soil data that can be efficiently acquired remotely and proximally, and freely available open-source algorithms, have led to an accelerated adoption of ML techniques to analyse soil data. Given the large number of publications, it is an impossible task to manually review all papers on the application of ML in soil science without narrowing down a narrative of ML application in a specific research question. This paper aims to provide a comprehensive review of the application of ML techniques in soil science aided by a ML algorithm (Latent Dirichlet Allocation) to find patterns in a large collection of text corpus. The objective is to gain insight into publications of ML applications in soil science and to discuss the research gaps in this topic. We found that: a) there is an increasing usage of ML methods in soil sciences, mostly concentrated in developed countries, b) the reviewed publication can be grouped into 12 topics, namely remote sensing, soil organic carbon, water, contamination, methods (ensembles), erosion and parent material, methods (NN, SVM), spectroscopy, modelling (classes), crops, physical and modelling (continuous), c) advanced ML methods usually perform better than simpler approaches thanks to their capability to capture non-linear relationships. From these findings, we found research gaps, in particular: about the precautions that should be taken (parsimony) to avoid overfitting, and that the interpretability of the ML models is an important aspect to consider when applying advanced ML methods in order to improve our knowledge and understanding of soil. We foresee that a large number of studies will focus on the latter topic.

Description: In this work, we provide a preliminary assessment of whether or not ramped thermal oxidation coupled with determination of the radiocarbon content of the evolved CO2 can be used to isolate biologically meaningful fractions of SOM along with direct information on the turnover rate of each fraction. Using a 30 year time-series of soil samples from a well characterized agronomic trial, we found that the incorporation of the bomb-spike in atmospheric 14CO2 into thermal fractions could be successfully modelled. With increasing activation energy of the fraction, the mean residence time of the fraction increased from 10 to 400 years. Importantly, the first four of five thermal fractions appeared to be a mixture of fast and increasingly slower cycling SOM. To further understand the composition of different thermal fractions, stepped pyrolysis-gas chromatography-mass spectrometry (py-GC/MS) experiments were performed at five temperatures ranging from 330 to 735 °C. The py-GC/MS data showed a reproducible shift in chemistry across the temperature gradient trending from polysaccharides and lipids at low temperature to lignin and microbial-derived compounds at middle temperatures to aromatic and unknown compounds at the highest temperatures. Integrating the 14C and Py-GC-MS data suggests the organic compounds, with the exception of aromatic moieties likely derived from wildfire, with centennial residence times are not more complex but may be protected from pyrolysis, and likely also from biological mineralization, by interactions with mineral surfaces.

Description: Four purified phytases isolated from Aspergillus niger and Escherichia coli were characterized biochemically and in terms of their adsorption to soils belonging to the Mollisol order. Three different organic P substrates were used to measure enzyme activity in a wide range of pH (2.3 to 9) and temperatures (−10° to 70 °C): p-nitrophenyl-phosphate, glyceraldehyde-3-phosphate and phytic acid. Phytases from A. niger showed a higher capacity to release P (36 to 50 % of P contained in the substrates, 44 to 62 μg P), than phytases from E. coli (24 to 36 %, 20 to 44 μg P). The amount of P released from organic P substrates by A. niger phytases followed the following range: p-nitrophenyl-phosphate 〉 glyceraldehyde-3-phosphate 〉 phytic acid whereas in E. coli phytases the order was p-nitrophenyl-phosphate/glyceraldehyde-3-phosphate 〉 phytic acid. All phytases were active throughout the pH and temperature ranges for optimum crop production. The proportion of phytases found in the solid phase of the soil 60 minutes after addition was lower than that found in the liquid phase (23–34 % vs. 66–77 %). Obtained results are promising in terms of the use of phytases as a complement to P fertilization in agricultural settings and encourages further studies under field conditions.

Description: The Fukushima Dai-ichi Nuclear Power Plant (FDNPP) accident in March 2011 resulted in the contamination of Japanese landscapes with radioactive fallout. Accordingly, the Japanese authorities decided to conduct extensive remediation activities in the impacted region to allow for the relatively rapid return of the local population. The objective of this review is to provide an overview of the decontamination strategies and their potential effectiveness in Japan, focussing on particle-bound radiocesium. In the Fukushima Prefecture, the decision was taken to decontaminate the fallout-impacted landscapes in November 2011 for the 11 municipalities evacuated after the accident (Special Decontamination Zones – SDZ, 1117 km2) and for the 40 non-evacuated municipalities affected by lower, although still significant, levels of radioactivity (Intensive Contamination Survey Areas, 7836 km2). Decontamination activities predominantly targeted agricultural landscapes and residential areas. No decontamination activities are currently planned for the majority of forested areas, which cover ~ 75 % of the main fallout-impacted region. Research investigating the effectiveness of decontamination activities underlined the need to undertake concerted actions at the catchment scale to avoid the renewed supply of contamination from the catchment headwaters after the completion of remediation activities. Although the impact of decontamination on the radioactive dose rates for the local population remains a subject of debate in the literature and in the local communities, outdoor workers in the SDZ represent a group of the local population that may exceed the long-term dosimetric target of 1 mSv yr−1. Decontamination activities generated ~ 20 million m3 of soil waste by early 2019. The volume of waste generated by decontamination may be decreased through incineration of combustible material and recycling of the less contaminated soil for civil engineering structures. However, most of this material will have to be stored for ~ 30 years at interim facilities opened in 2017 in the close vicinity of the FDNPP before being potentially transported to final disposal sites outside of the Fukushima Prefecture. Further research is required to investigate the perennial contribution of radiocesium from forest sources. In addition, the re-cultivation of farmland after decontamination raises additional questions associated with the fertility of remediated soils and the potential transfer of residual radiocesium to the plants. Overall, we believe it is important to synthesize the remediation lessons learnt following the FDNPP nuclear accident, which could be fundamental if a similar catastrophe occurs somewhere on Earth in the future.

Description: Net loss of soil organic carbon (SOC) from terrestrial ecosystems is a likely consequence of global warming and this may affect key soil functions. Strongest changes in temperature are expected to occur at high northern latitudes, with boreal forest and tundra as prevailing land-cover types. However, specific ecosystem responses to warming are understudied. We used a natural geothermal soil warming gradient in an Icelandic spruce forest (0–17.5 °C warming intensity) to assess changes in SOC content in 0–10 cm (topsoil) and 20–30 cm (subsoil) after 10 years of soil warming. Five different SOC fractions were isolated and the amount of stable aggregates (63–2000 µm) was assessed to link SOC to soil structure changes. Results were compared to an adjacent, previously investigated warmed grassland. Soil warming had depleted SOC in the forest soil by −2.7 g kg−1 °C−1 (−3.6 % °C−1) in the topsoil and −1.6 g kg−1 °C−1 (−4.5 % °C−1) in the subsoil. Distribution of SOC in different fractions was significantly altered, with particulate organic matter and SOC in sand and stable aggregates being relatively depleted and SOC attached to silt and clay being relatively enriched in warmed soils. The major reason for this shift was aggregate break-down: topsoil aggregate mass proportion was reduced from 60.7 ± 2.2 % in the unwarmed reference to 28.9 ± 4.6 % in the most warmed soil. Across both depths, loss of one unit SOC caused a depletion of 4.5 units aggregated soil, which strongly affected bulk density (R2 = 0.91 when correlated to SOC and R2 = 0.51 when correlated to soil mass in stable aggregates). The proportion of water extractable carbon increased with decreasing aggregation, indicating an indirect SOC protective effect of aggregates 〉 63 µm. Topsoil changes in total SOC and fraction distribution were more pronounced in the forest than in the adjacent warmed grassland soils, due to higher and more labile initial SOC. However, no ecosystem effect was observed in the response of subsoil SOC and fraction distribution. Whole profile differences across ecosystems might thus be small. Changes in soil structure upon warming should be studied more deeply and taken into consideration when interpreting or modelling biotic responses to warming.

Description: Recent studies indicate that climate change influences soil mineralogy by altering weathering processes, and thus impacts soil aggregation and organic carbon (SOC) stability. Alpine ecosystems of the Neotropical Andes are characterized by high SOC stocks, which are important to sustain ecosystem services. However, climate change in the form of altered precipitation patterns can potentially affect soil aggregation and SOC stability with potentially significant effects on the soil’s ecosystem services. This study aimed to investigate the effects of precipitation and lithology on soil aggregation and SOC stability in the Peruvian Andean grasslands, and assessed whether occlusion of organic matter (OM) in aggregates controls SOC stability. For this, samples were collected from limestone soils (LSs) and acid igneous rock soils (ASs) from two sites with contrasting precipitation levels. We used a dry-sieving method to quantify aggregate size distribution, and applied a 76-day soil incubation with intact and crushed aggregates to investigate SOC stability in dependence on aggregation. SOC stocks ranged from 153±27 to 405 ± 42 Mg ha−1, and the highest stocks were found in the LSs of the wet site. We found lithology rather than precipitation to be the key factor regulating soil aggregate size distribution, as indicated by coarse aggregates in the LSs and fine aggregates in the ASs. SOC stability estimated by specific SOC mineralization rates decreased with precipitation in the LSs, but minor differences were found between wet and dry sites in the ASs. Aggregate destruction had a limited effect on SOC mineralization, which indicates that occlusion of OM in aggregates played a minor role in OM stabilization. This was further supported by inconsistent patterns of aggregate size distribution compared to the patterns of SOC stability. We propose that OM adsorption on mineral surfaces is the major OM stabilization mechanism controlling SOC stocks and stability. The results highlight the interactions between precipitation and lithology on SOC stability, which are likely controlled by soil mineralogy in relation to OM input.

Description: Since decades, microplastics and microglass enter aquatic and terrestrial environments. The complexity of the environmental impact is difficult to capture and consequences on ecosystem components e.g. such as soil microorganisms are virtually unknown. Addressing this issue, we performed an incubation experiment by adding 1 % of five different types of impurities (≤ 100 µm) to an agricultural used soil (Chernozem). Four microplastic types (polypropylene (PP), low density polyethylene (LD-PE), polystyrene (PS) and polyamide12 (PA12)) and microglass were used as treatment variants. After 80 days of incubation at 20 °C, we examined soil microbial community structure by using phospholipid fatty acids (PLFA) as markers for bacteria, fungi and protozoa. The results showed that soil microorganisms were not significantly affected by the presence of microplastic and microglass. However, PLFAs tend to increase in LD-PE (27 %), PP (18 %) and microglass (11 %) treated soil in comparison with untreated soil, whereas PLFAs in PA12 (32 %) and PS (11 %) treated soil decreased. Interestingly, the comparison of PLFA contents between microplastic types revealed significant differences of PA12 (−87 %) and PS (−42 %) compared to LD-PE. Furthermore, bacterial PLFAs showed a much higher variability after microplastic incubation whereby fungi seem to be more unaffected after 80 days of incubation. Same for protozoa, which were more or less unaffected by microplastic treatment showing only minor reduction of the PLFA contents compared to control. In contrast, microglass has obviously an inhibiting effect on protozoa because PLFAs were under the limit of determination. Our study provides hints, that microplastics have, depending on type, contrary effects on soil microbiology and microglass seems to be highly toxic for protozoa.

Description: The activity of microorganisms in soil is important for a robust functioning soil and related ecosystem service. Hence, there is a necessity to identify the indigenous soil microbial community for its functional properties using soil microbiological methods in order to determine the natural properties, functioning and operating range of soil microbial communities, and to assess ecotoxicological effects due to anthropogenic activities. Numerous microbiological methods currently exist in the literature and new, more advanced methods continue to be developed; however, only a limited number of the methods are standardized. Consequently, there is a need to identify the most promising non-standardized methods for assessing soil quality and develop these into standards. In alignment with the "Ecosystem Service Approach", new methods should focus on soil microbial function, including nutrient cycling, pest control and plant growth promotion, carbon cycling and sequestration, greenhouse gas emission, and soil structure. The few existing, function-related standard methods available focus on the estimation of microbial biomass, basal respiration, enzyme activities related to nutrient cycling, and organic chemical biodegradation. This paper sets out to summarize and expand on recent discussions within the International Organization for Standardization (ISO), Soil Quality - Biological Characterization sub-committee (ISO TC 190/SC 4) where a need was identified to develop scientifically sound methods which would best fulfil the practical needs of future users for assessing soil quality. Of particular note was the current evolution of molecular methods in microbial ecology that uses quantitative real time PCR (qPCR) to produce a large number of new endpoints and is more sensitive as compared to "classical" methods. Quantitative PCR assesses the activity of microbial genes that code for enzymes that catalyse major transformation steps in nitrogen and phosphorus cycling, greenhouse gas emissions, chemical transformations including pesticide degradation, and plant growth promotion pathways. In the assessment of soil quality methods, it was found that fungal methods were significantly underrepresented. As such, techniques to analyse fungal enzyme activities are proposed. Additionally, methods for the determination of microbial growth rates and efficiencies, including the use of glomalin as a biochemical marker for soil aggregation, are discussed. Furthermore, field methods indicative of carbon turnover, including the litter bag test and a modification to the tea bag test, are presented. As a final note, it is suggested that endpoints should represent a potential function of soil microorganisms rather than actual activity levels, as the latter can largely be dependent on short-term variable soil properties such as pedoclimatic conditions, nutrient availability, and anthropogenic soil cultivation activities.