Beyond prediction: methods for interpreting complex models of soil variation

Understanding the spatial variation of soil properties is central to many sub-disciplines of soil science. Commonly in soil mapping studies, a soil map is constructed through prediction by a statistical or non-statistical model calibrated with measured values of the soil property and environmental covariates of which maps are available. In recent years, the field has gradually shifted attention towards more complex statistical and algorithmic tools from the field of machine learning. These models are particularly useful for their predictive capabilities and are often more accurate than classical models, but they lack interpretability and their functioning cannot be readily visualized. There is a need to understand how these these models can be used for purposes other than making accurate prediction and whether it is possible to extract information on the relationships among variables found by the models. In this paper we describe and evaluate a set of methods for the interpretation of complex models of soil variation. An overview is presented of how model-independent methods can serve the purpose of interpreting and visualizing different aspects of the model. We illustrate the methods with the interpretation of two mapping models in a case study mapping topsoil organic carbon in France. We reveal the importance of each driver of soil variation, their interaction, as well as the functional form of the association between environmental covariate and the soil property. Interpretation is also conducted locally for an area and two spatial locations with distinct land use and climate. We show that in all cases important insights can be obtained, both into the overall model functioning and into the decision made by the model for a prediction at a location. This underpins the importance of going beyond accurate prediction in soil mapping studies. Interpretation of mapping models reveal how the predictions are made and can help us formulating hypotheses on the underlying soil processes and mechanisms driving soil variation.

Do forest over- and understory respond to the same environmental variables when viewed at the taxonomic and trait level?

While the relative importance of climate filtering is known to be higher for woody species assemblages than herbaceous assemblage, it remains largely unexplored whether this pattern is also reflected between the woody overstory and herbaceous understory of forests. While climatic variation will be more buffered by the tree layer, the understory might also respond more to small-scale soil variation, next to experiencing additional environmental filtering due to the overstory's effects on light and litter quality. For (sub)tropical forests, the understory often contains a high proportion of fern and lycophyte species, for which environmental filtering is even less well understood. We explored the proportional importance of climate proxies and soil variation on the species, functional trait and (functional) diversity patterns of both the forest overstory and fern and lycophyte understory along an elevational gradient from 850 to 2100 m a.s.l. in northern Taiwan. We selected nine functional traits expected to respond to soil nutrient or climatic stress for this study and furthermore verified whether they were positively related across vegetation layers, as expected when driven by similar environmental drivers. We found that climate was a proportionally more important predictor than soil for the species composition of both vegetation layers and trait composition of the understory. The stronger than expected proportional effect of climate for the understory was likely due to fern and lycophytes' higher vulnerability to drought, while the high importance of soil for the overstory seemed driven by deciduous species. The environmental drivers affected different response traits in both vegetation layers, however, which together with additional overstory effects on understory traits, resulted in a strong disconnection of community-level trait values across layers. Interestingly, species and functional diversity patterns could be almost exclusively explained by climate effects for both vegetational layers, with the exception of understory species richness. This study illustrates that environmental filtering can differentially affect species, trait and diversity patterns and can be highly divergent for forest overstory and understory vegetation, and should consequently not be extrapolated across vegetation layers or between composition and diversity patterns.

Introduction to Precision Agriculture

The global population is increasing at a tremendous speed; thus, the demand for safe and secure food to meet this population is in demand. Therefore, traditional farming methods are insufficient to meet this demand; thus, the next revolution in agriculture is required, which is Precision Agriculture (PA), the Fourth Agriculture Revolution. PA is a technology where the concept of farm management is based on observation, measuring, and responding to inter- and intra-field variability in crops. The technologies used for performing precision agriculture are mapping, global positioning system (GPS), yield monitoring and mapping, grid soil sampling application, variable-rate fertilizer application, remote sensing, geographic information systems (GIS), quantifying on farm variability, soil variation, variability of soil water content, time and space scales, robots, drones, satellite imagery, the internet of things, smartphone, and machine learning. Hence, the current chapter will be emphasizing the overview, concepts, history, world interest, benefits, disadvantages, and precision farming needs.

Management Zone Delineation for Site-Specific Fertilization in Rice Crop Using Multi-Temporal RapidEye Imagery

The objective of this research is to assess the potential of satellite imagery in detecting soil heterogeneity, with a focus on site-specific fertilization in rice. The basic hypothesis is that spectral variation would express soil fertility variations analogously. A 100-ha rice crop, located in the Plain of Thessaloniki, Greece, was selected as the study area for the 2016 cropping season. Three RapidEye images were acquired during critical growth stages of rice cultivation from the previous year (2015). Management zones were delineated with image segmentation of a 15-band multi-temporal composite of the RapidEye images (three dates × five bands), using the Fractal Net Evolution Approach (FNEA) algorithm. Then, an equal number of soil samples were collected from the centroid of each management zone before seedbed preparation. The between-zone variation of the soil properties was found to be 33.7% on average, whereas the within-zone variation 18.2%. The basic hypothesis was confirmed, and moreover, it was proved that zonal applications reduced within-zone soil variation by 18.6% compared to conventional uniform applications. Finally, between-zone soil variation was significant enough to dictate differentiated fertilization recommendations per management zone by 24.5% for the usual inputs.

pH effect on strain-specific transcriptomes of the take-all fungus

The soilborne fungus Gaeumannomyces graminis var. tritici (Ggt) causes the take-all disease on wheat roots. Ambient pH has been shown to be critical in different steps of Ggt life cycle such as survival in bulk soil, saprophytic growth, and pathogenicity on plants. There are however intra-specific variations and we previously found two types of Ggt strains that grow preferentially either at acidic pH or at neutral/alkaline pH; gene expression involved in pH-signal transduction pathway and pathogenesis was differentially regulated in two strains representative of these types. To go deeper in the description of the genetic pathways and the understanding of this adaptative mechanism, transcriptome sequencing was achieved on two strains (PG6 and PG38) which displayed opposite growth profiles in two pH conditions (acidic and neutral). PG6, growing better at acidic pH, overexpressed in this condition genes related to energy production and protein deubiquitination. In contrast, PG38, which grew better at neutral pH, overexpressed in this condition genes involved in fatty acids metabolism. This strain also expressed stress resistance mechanisms at both pH, to assert a convenient growth under various ambient pH conditions. These differences in metabolic pathway expression between strains at different pH might buffer the effect of field or soil variation in wheat fields, and explain the success of the pathogen.

Assessment of reclaimed soils by unsupervised clustering of proximal sensor data

The application of soil proximal sensors on reclaimed sites presents a novel method for assessing the quality of reclaimed landscapes. This method improves assessment reliability, information management, and environmental assurance. One proximal sensing system that could be used to provide high spatial resolution measurements of soil parameters is an on-the-go optical sensor that collects data at two wavelengths: 660 and 940 nm. Proximal soil sensing data were collected at 27 sites, where organic matter, cation exchange capacity (CEC), and soil water content were collected from 221 soil samples from 0 to 15 cm. The proximal soil sensor data were then automatically clustered using a combination of self-organizing maps and random uniform forests. Overall, the proximal sensor data combined with this data analysis approach created maps with either three or four soil zones. On average, soil zones had statistically significant differences in organic matter, CEC, and water content. This system could be used to map out zones with significant soil variation as part of reclamation monitoring and then used to guide laboratory analytical sampling. Future work should focus on development of on-the-go reflectance spectroscopy systems to provide quantitative soil data with high spatial resolution.