Deep learning and process understanding for data-driven Earth system science

Nature, Journal Year: 2019, Volume and Issue: 566(7743), P. 195 - 204

Published: Feb. 1, 2019

Language: Английский

---

A global atlas of the dominant bacteria found in soil

Science, Journal Year: 2018, Volume and Issue: 359(6373), P. 320 - 325

Published: Jan. 18, 2018

The immense diversity of soil bacterial communities has stymied efforts to characterize individual taxa and document their global distributions. We analyzed soils from 237 locations across six continents found that only 2% phylotypes (~500 phylotypes) consistently accounted for almost half the worldwide. Despite overwhelming communities, relatively few are abundant in globally. clustered these dominant into ecological groups build first atlas taxa. Our study narrows down number a "most wanted" list will be fruitful targets genomic cultivation-based aimed at improving our understanding microbes contributions ecosystem functioning.

Language: Английский

---

The global tree restoration potential

Science, Journal Year: 2019, Volume and Issue: 365(6448), P. 76 - 79

Published: July 4, 2019

The restoration of trees remains among the most effective strategies for climate change mitigation. We mapped global potential tree coverage to show that 4.4 billion hectares canopy cover could exist under current climate. Excluding existing and agricultural urban areas, we found there is room an extra 0.9 cover, which store 205 gigatonnes carbon in areas would naturally support woodlands forests. This highlights as our solution date. However, will alter this coverage. estimate if cannot deviate from trajectory, may shrink by ~223 million 2050, with vast majority losses occurring tropics. Our results highlight opportunity mitigation through but also urgent need action.

Language: Английский

---

Soil organic carbon storage as a key function of soils - A review of drivers and indicators at various scales

Geoderma, Journal Year: 2018, Volume and Issue: 333, P. 149 - 162

Published: July 23, 2018

Language: Английский

---

SoilGrids 2.0: producing soil information for the globe with quantified spatial uncertainty

SOIL, Journal Year: 2021, Volume and Issue: 7(1), P. 217 - 240

Published: June 14, 2021

Abstract. SoilGrids produces maps of soil properties for the entire globe at medium spatial resolution (250 m cell size) using state-of-the-art machine learning methods to generate necessary models. It takes as inputs observations from about 240 000 locations worldwide and over 400 global environmental covariates describing vegetation, terrain morphology, climate, geology hydrology. The aim this work was production properties, with cross-validation, hyper-parameter selection quantification spatially explicit uncertainty, implemented in version 2.0 product incorporating practices adapting them digital mapping legacy data. paper presents evaluation predictions produced organic carbon content, total nitrogen, coarse fragments, pH (water), cation exchange capacity, bulk density texture fractions six standard depths (up 200 cm). quantitative showed metrics line previous global, continental large-region studies. qualitative that coarse-scale patterns are well reproduced. uncertainty scale highlighted need more observations, especially high-latitude regions.

Language: Английский

---

Rainfall–runoff modelling using Long Short-Term Memory (LSTM) networks

Hydrology and earth system sciences, Journal Year: 2018, Volume and Issue: 22(11), P. 6005 - 6022

Published: Nov. 22, 2018

Abstract. Rainfall–runoff modelling is one of the key challenges in field hydrology. Various approaches exist, ranging from physically based over conceptual to fully data-driven models. In this paper, we propose a novel approach, using Long Short-Term Memory (LSTM) network, special type recurrent neural network. The advantage LSTM its ability learn long-term dependencies between provided input and output which are essential for storage effects e.g. catchments with snow influence. We use 241 freely available CAMELS data set test our approach also compare results well-known Sacramento Soil Moisture Accounting Model (SAC-SMA) coupled Snow-17 routine. show potential as regional hydrological model predicts discharge variety catchments. last experiment, possibility transfer process understanding, learned at scale, individual thereby increasing performance when compared trained only on single Using were able achieve better SAC-SMA + Snow-17, underlines applications.

Language: Английский

---

Soil carbon debt of 12,000 years of human land use

Proceedings of the National Academy of Sciences, Journal Year: 2017, Volume and Issue: 114(36), P. 9575 - 9580

Published: Aug. 21, 2017

Significance Land use and land cover change has resulted in substantial losses of carbon from soils globally, but credible estimates how much soil been lost have difficult to generate. Using a data-driven statistical model the History Database Global Environment v3.2 historic land-use dataset, we estimated that agricultural uses loss 133 Pg C soil. Importantly, our maps indicate hotspots loss, often associated with major cropping regions degraded grazing lands, suggesting there are identifiable should be targets for restoration efforts.

Language: Английский

---

Global threat of arsenic in groundwater

Science, Journal Year: 2020, Volume and Issue: 368(6493), P. 845 - 850

Published: May 21, 2020

Dowsing for danger Arsenic is a metabolic poison that present in minute quantities most rock materials and, under certain natural conditions, can accumulate aquifers and cause adverse health effects. Podgorski Berg used measurements of arsenic groundwater from ∼80 previous studies to train machine-learning model with globally continuous predictor variables, including climate, soil, topography (see the Perspective by Zheng). The output global map reveals potential hazard contamination groundwater, even many places where there are sparse or no reported measurements. highest-risk regions include areas southern central Asia South America. Understanding especially essential facing current future water insecurity. Science , this issue p. 845 ; see also 818

Language: Английский

---

Soil nematode abundance and functional group composition at a global scale

Nature, Journal Year: 2019, Volume and Issue: 572(7768), P. 194 - 198

Published: July 24, 2019

Language: Английский

---

The Ecology of Soil Carbon: Pools, Vulnerabilities, and Biotic and Abiotic Controls

Annual Review of Ecology Evolution and Systematics, Journal Year: 2017, Volume and Issue: 48(1), P. 419 - 445

Published: Sept. 6, 2017

Soil organic matter (SOM) anchors global terrestrial productivity and food fiber supply. SOM retains water soil nutrients stores more carbon than do plants the atmosphere combined. is also decomposed by microbes, returning CO 2 , a greenhouse gas, to atmosphere. Unfortunately, stocks have been widely lost or degraded through land use changes unsustainable forest agricultural practices. To understand its structure function maintain restore SOM, we need better appreciation of (SOC) saturation capacity retention above- belowground inputs in SOM. Our analysis suggests root are approximately five times likely an equivalent mass aboveground litter be stabilized as Microbes, particularly fungi bacteria, faunal webs strongly influence decomposition at shallower depths, whereas mineral associations drive stabilization depths greater ∼30 cm. Global uncertainties amounts locations include extent wetland, peatland, permafrost systems factors that constrain such shallow bedrock. In consideration these uncertainties, estimate SOC 3 m between 2,270 2,770 Pg, respectively, but could much 700 Pg smaller. Sedimentary deposits deeper contain >500 additional SOC. Soils hold largest biogeochemically active pool on Earth critical for stabilizing atmospheric concentrations. Nonetheless, pressures soils continue from management, including increasing bioenergy production.

Language: Английский

---