the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Oct 2023
The soil knowledge library (KLIB) – a structured literature database on soil process research
Hans-Jörg Vogel
Bibiana Betancur-Corredor
Leonard Franke
Sara König
Birgit Lang
Maik Lucas
Eva Rabot
Bastian Stößel
Ulrich Weller
Martin Wiesmeier
Ute Wollschläger
In this technical note, we introduce a web-based application, the BonaRes Knowledge Library (KLIB, https://klibrary.bonares.de, last access: 26 July 2023), for the compilation and classification of scientific publications on soil processes according to the specific site conditions and experimental boundary conditions. The tool was developed based on the understanding that experimental findings in soil science are highly dependent on soil type, land use, and climate. The KLIB, therefore, goes beyond other available digital libraries by providing metadata on the site conditions and experimental settings for each publication. A number of visualization tools have been developed in the form of graphical networks to illustrate, for example, publications sharing the same type of scientific questions or soil properties that are affected by different types of drivers. This should help to explore the contents of the literature database more efficiently in order to support and facilitate the literature search efforts of the users. The KLIB is designed as a collaborative effort to encourage soil scientists to participate by entering their own studies and extending the existing database.
- Article
(2620 KB) - Full-text XML
- BibTeX
- EndNote
Soil is a highly complex system where a multitude of physical, chemical, and biological processes interact to generate several essential soil functions such as fertility for plant growth, water storage and water purification, nutrient cycling, carbon storage, and habitats for biodiversity. Today, we are still far from a solid mechanistic understanding of all these interacting processes. However, as a fundamental principle of science, identifying characteristic patterns of observed phenomena is a highly valuable initial step towards an in-depth process-based understanding of how soil as a complex system functions.
One example is the repeated observation that rather old and stable organic carbon in soil is frequently of microbiological origin (Miltner et al., 2012). This led to the insight that stabilizing carbon in soil is not mainly a result of chemical recalcitrance as previously thought but rather the consequence of physical protection within the heterogeneous soil structure, as was already suspected by Schmidt et al. (2011).
Over the last 2 decades, scientific tools were developed with the intention of providing more solid grounds for the observations of empirical relations. One approach is the systematic review of the available literature, including a quantitative synthesis of research results. This approach, termed meta-analysis, was initially developed in medical research to synthesize data from multiple clinical trials (Koricheva et al., 2013). However, there are a number of critical issues that come along with meta-analysis – for example, the bias resulting from a deliberate selection of articles to synthesize and the publication bias reflecting the propensity for journals to publish studies with positive, hypothesis-affirming results (Haddaway et al., 2015). These sources of bias also apply to meta-analyses in soil science, with an additional aspect that is especially critical in this field. In the medical field, the object of interest, the human organism, operates in fundamentally the same way. In soil science, however, different soil types under different land uses and climates can be considered to be “organisms” that have quite different internal controls and process interactions. Also, the temporal and spatial scale of experiments can have a substantial impact on experimental results. Hence, the integration of studies carried out at different locations under different boundary conditions might lead to inconclusive, contrasting, or even misleading results.
This well-known problem in meta-analyses leads to “comparing apples and oranges”. It might be one of the reasons why different studies on the same subject come up with contradictory findings. For example, the impact of reduced tillage on the carbon level of soils was found to have a positive or negative effect or no effect at all (Balesdent et al., 2000; Simpson et al., 2023). Earthworm activity may stimulate organic matter mineralization (Wachendorf et al., 2014), or it may enhance carbon stabilization (Guggenberger et al., 1996). In this case, the difference between the studies lies in the timescale considered. Earthworms foster microbial activity in their guts and within the fresh casts, leading to increased mineralization rates. By contrast, in the long run, the organic matter within the relatively dense casts is stabilized over time since the accessibility for microbes is reduced. As another example, Jarvis et al. (2013) demonstrated how the saturated hydraulic conductivity depends on site conditions, especially land use, but not very much on soil texture. As is obvious from these examples, studies that are included in a meta-analysis need to be analyzed separately through meta-regressions that account for different site characteristics and experimental boundary conditions.
Identifying characteristic patterns of observed phenomena is certainly an important initial step towards an in-depth process-based understanding of how soil as a complex system functions. It is also required for process-based, predictive modeling of how changing boundary conditions (e.g., in terms of land use or climate) will impact the multitude of soil functions. Such a model would be a major advance in soil science. Currently, we mostly rely on empirical relations while our understanding of the relevant processes and interactions providing causality for the observed correlations is still limited.
In today's digital databases of scientific literature, which are accessible through various search engines, it is not yet possible to stratify the literature search according to soil types, site characteristics, or experimental boundary conditions. In addition, the search for specific processes and process interactions is difficult. The basic idea of the BonaRes Knowledge Library (KLIB) presented in this article is to allow for an analysis of the included research publications which can be structured according to site conditions, the investigated processes, and the main results of the studies. It is mainly focused on soil process research, but methodological publications might also be included. For process-oriented publications, all the metadata need to be entered for each individual study, which is a considerable effort. For this reason, KLIB is intended to be a community effort, in which many interested scientists can contribute with the idea that the work is shared.
In the following, we introduce the KLIB as an open-access tool for a structured analysis of published work on soil processes. We present examples of how to use KLIB in the literature search and in screening for meta-analysis, and we discuss the potential of the KLIB for analyzing published findings in general and how it might be helpful to identify research gaps.
2.1 Intention and functionality
The KLIB is a web-based tool accessible by any web browser (https://klibrary.bonares.de). It provides an interface for uploading relevant metadata related to scientific publications with a focus on soil process research. The main intention in developing this tool was the insight that observed soil behavior and the derived process understanding is highly sensitive to the boundary conditions under which the individual studies were conducted. The original idea for this tool emerged from the need to base the process understanding required for a mechanistic soil model on published knowledge. There are mainly two types of essential boundary conditions for any experimental study: first, the specific site characteristics (in terms of soil type, soil texture, land use, and climatic region) and second, the way observations and measurements were made (more specifically, at which time and spatial scales, in the field or in the laboratory, and using disturbed or undisturbed samples). The KLIB is a tool to record all the information for individual publications that we think is crucial for a clear interpretation of the results.
An additional but important category of information to be entered into the KLIB is the research question being investigated in the publication, the metrics or measured soil characteristics used to investigate that question, and the key results that have been obtained from it. Entering all this information involves a considerable amount of work. However, the additional effort is not huge. Moreover, this needs to be done only once for a given publication, the results can be used by the entire scientific community, and everyone can contribute – so, ideally, the work is shared among many colleagues.
The KLIB has a clear focus on publications reporting on experimental results that are suitable to improve our understanding of soil processes. Typical examples are as follows: how do different types of tillage practices affect the physical, biological, or chemical properties of soils, and what site factors are important in this context? Or how does a changing climate in terms of temperature and precipitation pattern impact the carbon balance in soils?
As the number of registered publications increases with time, the KLIB becomes an increasingly valuable tool for the structured analysis of published results. To this end, an essential additional functionality of the KLIB is the provision of data filters and visualization tools for a targeted evaluation of the database. As an example, all publications that deal with the impact of zero tillage on physical, chemical, and biological soil properties can be visualized in an interactive node graph panel. Because of the large number of publications, this graph can look considerably overloaded. Hence, the various visualizations include tools that enable the results to be filtered, e.g., by focusing on the comparison of conventional and zero tillage and looking at specific soil attributes such as soil carbon or bulk density. In this way, the contents of the library can be explored more efficiently. With the described functionality, the KLIB might improve the efficiency with which the scientific community can benefit from the continuously increasing number of publications.
2.2 Technical implementation
The KLIB is a web-based application that follows modern standards and techniques. For the front end, we use the framework Angular as a foundation for all continuing functions (Angular, 2023). The backend is built on node.js with Express.js to handle all the data structures and the connection to the database. While being visible and searchable for all users, all entered papers are connected to a unique login of each person, which gives editing rights only to authors and those who have been given access. The process of entering bibliographic data is supported by the option to extract the DOI from a PDF and retrieve the data from the Crossref-API. This allows us to get all information in seconds without the need to enter it manually.
To ensure security, we use the newest standards and techniques like JSON Web Tokens for authenticating to the KLIB, and on backend site, we use MongoDB to store the data secured. The entire database is backed up several times a day so that, in the event of a problem, all data entered can be restored to the most recent version possible. MongoDB provides the speed to prepare the data for complex visualizations in an acceptable amount of time. The JSON format is commonly used to store data and to send it through several API calls to the user on the frontend side.
We implemented complex algorithms to generate the information for the visualizations and to build the different networks that can be explored and adjusted. The presentation itself is handled by the D3.js library, which provides powerful tools to generate smooth networks. The implementation of canvas as the rendering structure is essential for handling big datasets and many points in a network; the conventionally used svg is too slow for our application and causes performance problems on many devices.
The whole structure of our KLIB is built to be easily accessible, and a framework like Angular keeps the whole structure modular and easy to update. Any new features or extra visualizations can be edited or implemented according to the feedbacks of the users.
2.3 Input of metadata
The bibliographic data for each publication, including the abstracts, are automatically extracted from the publication's PDF file by dropping the icon of the file into the KLIB web application. In some cases when this is not possible, e.g., if the PDF document of a publication was created by scanning a paper document into a bitmap format, these data can also be entered manually or extracted from a corresponding BibTeX file. For legal reasons, the original PDF file cannot be stored or made available in the KLIB. Afterward, the required metadata on site characteristics, sampling design, and key findings can be entered with the help of an intuitive user interface (Fig. 1). For experimental research articles, which are the main focus of the KLIB, a number of different categories of metadata can be provided (if available in the publication):
- a.
Type of study. Is it an experimental study, a model application, a methodological study, or a meta-analysis? And in the case of an experiment, does the study refer to individual sites or multiple sites? Is it a field or a lab experiment? What is the considered time scale?
- b.
Soil & Site. What are the soil types (with different classification schemes for selection available – e.g., WRB, USDA, FAO, German) and their texture? Which depths are investigated? Additional soil properties such as pH, organic carbon content, and bulk density can be provided optionally. What is the geographic region or spatial coordinates? What are climate characteristics (mean temperature, precipitation – these can also be derived from the specified location)? What is the land use?
- c.
Investigated Drivers. In case the effect of different treatments in terms of soil management or physico-chemical boundary conditions is investigated, these treatments, considered to be drivers, can be entered here.
- d.
Measured Variables. Which soil properties are measured and/or monitored?
- e.
Key Findings. Here, the relation between the drivers previously given with the monitored properties or the relation between different properties can be specified. These relations are provided in terms of simple effect categories (positive, negative, none, type of non-linear relations).
Figure 1Web interface of KLIB to enter metadata on site characteristics, experimental design, and key outcomes reported in a publication on soil process research. The required input is divided into the following categories: Type of Study, Site & Soil, Management & Land use, Measured Variables, and Key Findings (screenshot taken 18 July 2023).
Almost all entries can be selected via a structured tree of predefined and standardized terms following the AGROVOC thesaurus (Subirats-Coll et al., 2022). They were internally stored in a hierarchical tree of keywords which allows for an efficient exploration of the database. The entries can be chosen from different hierarchical levels depending on the availability of information. For example, it is possible to choose pH or, at the following level, either pH(water), pH(KCl), or pH(CaCl2). In visualization mode, the search for pH will show all the possible entries at the following level. The same principles apply to other entries in the hierarchical tree.
2.4 Exploring the database
All process-oriented publications included in the library are supplemented by metadata on different categories: investigated soil properties, soil management, and soil and site conditions. Besides a classical search for keywords, authors, and date of publication, the KLIB offers the functionality to explore its contents based on the interactive graphical tools that are designed to visualize complex interrelations within and among the different categories. Five different visualization tools are provided. They are described in the following section, alongside some typical application examples.
2.4.1 Keyword network (KN)
This visualization displays the network of publications that deal with a specified combination of soil properties or management practices. It is possible to combine an arbitrary number of keywords. A large number may produce an overloaded network, which may, however, depict the main focal points of the publications. An example is illustrated in Fig. 2a for a complex keyword network including physical and biological attributes together with all different tillage options. The two visual focal points in this graph are zero tillage and conventional tillage, which are obviously the two most frequently addressed topics in the current content of KLIB. A simple example is shown in Fig. 2b and c, where only the minimum number of two attributes was selected, namely bulk density and conservation tillage. The latter is a parent category in the keyword tree containing the four child nodes (i.e., ridge tillage, minimum tillage, zero tillage, and strip tillage) which are visualized as well. In this case, there were no publications in the database dealing with bulk density and strip tillage; hence, this point is missing in the visualized network. The publication metadata can be accessed by a click on the gray circles connecting pairs of attributes. The size of the circles corresponds to the number of publications that are provided as a list (Fig. 2d).
Figure 2Example of a complex keyword network including physical and biological attributes together with all different tillage options (a) (details not relevant here) and a simple network (c) where only bulk density and conservation tillage were selected (b). When clicking on the gray circles, the list of publications dealing with the connected attributes is shown (d) (screenshot taken 18 July 2023).
Figure 3Visualization mode “Properties affected by management” (PM) for bulk density (a) and earthworm abundance (b), showing which management options are compared with respect to their impact on the chosen soil property. When clicking on the gray circles, the corresponding list of publications, together with their key findings, is shown (screenshot taken 18 July 2023).
2.4.2 Properties affected by management (PM)
This visualization mode is focused on a single soil property such as soil organic carbon, bulk density, or earthworm abundance. It displays management practices that affect this specific property. The visualized network allows for the identification of publications that compare a set of different management practices with respect to the selected property. An example of bulk density in relation to tillage practices is presented in Fig. 3a, showing that the impacts of conventional tillage and zero tillage on bulk density are often compared to other tillage systems. Another example of earthworm abundance affected by different tillage practices is shown in Fig. 3b. By default, all relevant management practices are plotted, but this can be narrowed down to a convenient subset by the user. Again, the list of individual publications can be found with a click on the gray circles, where the key findings of each individual publication with respect to the chosen soil property can be found as well.
Figure 4Visualization mode “Drivers affecting properties” (DP). Plot (a) is the result for the selected driver of conventional tillage (in the center) when all other drivers which were investigated with respect to the same properties as conventional tillage are plotted. In plot (b) this graph is reduced to conventional tillage vs. zero tillage only. Plot (c) illustrates the soil properties affected by organic fertilizers vs. inorganic fertilizers. When clicking on the gray circles, the corresponding list of publications together with their key findings is shown (screenshot taken 18 July 2023).
2.4.3 Drivers affecting properties (DP)
This visualization mode is complementary to the PM mode. It focuses on drivers which can be either management practices, soil properties, or site conditions, and it displays which soil properties are affected by a chosen driver. In addition, the publications that report information about other drivers that were found to have an effect on the same properties are considered as well in the illustrated network. This can lead to rather complex networks, as shown in Fig. 4a for conventional tillage as the chosen driver. There are a large number of soil properties affected by conventional tillage, as well as a large number of other management practices that were investigated in comparison to conventional tillage. If this is reduced to only zero tillage, the structure of the graph becomes much less complex (Fig. 4b). In this way, references for the impact of different drivers can be identified, and their results can be compared. Another simple example is shown in Fig. 4c for the properties affected by organic fertilizers in comparison to inorganic ones. Of course, these visualizations might also become more complex as the number of publications included in the KLIB increases.
2.4.4 Property–property relation (PP)
This visualization mode provides relationships between measured soil properties. Such comparisons typically look for correlations between soil properties rather than causal relationships between a particular driver and selected soil properties, as in the previous visualization tools. An example is illustrated in Fig. 5a, where soil organic carbon was chosen by the user, and all other properties that have been investigated in combination with it are illustrated in the graph. The related publications are shown in gray circles. The categories of displayed properties can be reduced according to the interests of the user.
2.4.5 Publication-centered visualization (PC)
This visualization mode can be accessed for each individual publication. It illustrates all soil properties and drivers addressed in the given publication which is placed in the center of the graph (Fig. 5b). Other publications that share at least two of these properties and drivers are shown as well. The related publications are shown in the gray circles linked to those aspects. This visualization tool is designed to search for other publications dealing with closely related topics.
Figure 5Visualization mode “Property–property relation” (PP) displays all soil properties which were investigated together with some chosen property. Panel (a) shows an example of soil organic matter as the chosen property. The publication-centered visualization (b) displays all soil properties (orange circles) that are investigated in a chosen publication (red rectangle in the center) and provides all other publications that share at least two of the investigated properties of the chosen publication. In any case, the identified publications can be accessed through the gray circles connecting those properties. As an example, the publication list generated after choosing the gray circle marked by the green arrow is shown in Fig. 6 (screenshot taken 18 July 2023).
For any visualization mode, the results can be filtered with respect to soil and site conditions. It is possible to reduce the visualized results to specific soil properties like texture classes, pH ranges, or carbon contents or to site properties like climate zones.
Each visualization tool leads to a selection of items that match the criteria being searched for. In all visualization tools, this selection is reached by the gray circles in the graphics and is presented as a list of publications (Fig. 6). For each publication on this list, the key findings are provided through intuitive symbols (e.g., +, −, or x for positive, negative, or no relation, respectively). This corresponds to what was indicated as key findings when the publication was entered into the KLIB. Moreover, additional information on each publication is provided. The title of the publication is linked to the website of the publisher, and the symbol directly after the title provides a link to Google Scholar for this paper. By clicking on various intuitive symbols associated with each publication, the user can access all information stored for each publication. Alternatively, the publication-centered visualization mode can be invoked. Additional options are available to the user who originally entered the item into the library. The user can edit the information about this publication, remove the publication from the library, or share the rights to edit the content with other users. The latter can be very helpful when working in collaborative groups. An example is illustrated in (Fig. 6).
Figure 6Typical example for a list of publications dealing with the relationship between earthworm abundances and soil organic matter (Fig. 5a). The type of relationship is indicated for each individual publication and summarized for all publications on top (green arrows). The symbols behind each publication (red arrow) provide access to (i) the abstract and a summary of all metadata of the publication, (ii) the publication-centered visualization, and (iii) the download of all information related to this publication. For the user who has entered the publication into the library, there are additional options (blue arrow): (i) share the rights to edit with other users, (ii) add and edit comments to the publication or edit the publication information, or (iii) remove the publication from the library (screenshot taken 18 July 2023).
All search results can be downloaded in the open standard file format JSON, including all information stored for each individual publication.
2.5 Potential applications
The KLIB can be used in various ways, which are briefly outlined below.
2.5.1 Basic literature search
A basic mode of using the KLIB is for the search for publications dealing with a specific subject. Thereby, the search is not only reduced to keywords and their possible combinations. The search can be highly targeted to specific interactions between soil properties or to the effect of different drivers on soil properties. A special feature is direct access to the metadata on the site conditions for each publication. This allows the search for literature to be narrowed according to specific site conditions, such as soil type, climate zone, or vegetation.
A good example of a current research question is to search for published findings on the influence of reduced tillage on stabilized soil organic carbon. It is expected that some studies show a clear positive effect compared to conventional tillage, while others have found no effect or even a negative effect. The difference may lie in the site conditions or the specific experimental settings, such as soil type, initial carbon content, the timescale of the experiments, crop rotations studied, and climatic conditions. The KLIB provides the means to structure published findings according to relevant boundary conditions, as well as to find related literature unknown to the user so far.
2.5.2 Conducting systematic reviews and meta-analysis
The main objective of hypothesis-driven meta-analyses is to generate a quantitative estimate of a studied phenomenon, i.e., the effectiveness of an intervention, by gathering data from independent primary studies that focus on the same research question. This can only be accomplished by evidence consolidation from multiple studies that allow for side-by-side comparison and by filtering out the signal from the noise (Mikolajewicz and Komarova, 2019). For this purpose, the KLIB provides the necessary tools to stratify the search results along critical boundary conditions. An example would be a meta-analysis on the influence of earthworms on the saturated hydraulic conductivity of the soil. Based on a number of publications, a meta-analysis can be conducted to estimate the correlation between earthworm abundance and measured hydraulic conductivity. Considering that sandy soils with larger particle sizes typically have a much higher hydraulic conductivity than fine-textured soils, the results of the meta-analysis will likely lead to contradictory results if the results presented in the included publications are not analyzed separately for different soil textures. This is because, at the same time, sandy soils are often an unfavorable habitat for earthworms because of their low water-holding capacity and the low mechanical stability of the soil structure. Hence, a meta-analysis would be potentially biased by the results presented in studies conducted in sandy soils, which will likely show high hydraulic conductivity and low earthworm abundance. In this case, the conclusions with respect to the original research question could be completely wrong. The KLIB provides the tools to systematically detect potentially misleading results from the initial literature search and to find alternative ways to avoid them by providing information on critical boundary conditions.
The process of conducting a meta-analysis begins by formulating a question or defining the scope. The motivation can be to evaluate a theoretical hypothesis, focusing on causality, or to test for a consistent relationship between two or more variables, focusing on correlations. Therefore, it is crucial to avoid mixing studies that identify causal relationships through experimental manipulation with those that report observed relationships (Cote and Jennions, 2013). The KLIB can assist in distinguishing between causal and correlational studies since the user specifies the type of relationship between drivers and monitored properties for each publication.
In most cases for soil science research, the key finding of a meta-analysis is the mean effect of a certain intervention, whether it differs from the null expectation, and whether the variation of the outcome of the different studies can be explained. The question regarding the variation invites the meta-analyst to think about factors that differ among studies and that can affect the estimated mean effect. At this stage, the KLIB can help the meta-analyst to categorize or search for studies among different important groupings like site conditions, climate, management-related factors, type of study (field or laboratory), duration of the study, or methodology used, among many other factors.
Once the research questions are formulated, a protocol (i.e., following the PRISMA framework; Page et al., 2021) is needed for meta-analysis; this protocol will specify an objective search strategy, identify the inclusion criteria for the studies, and discuss potential sources of heterogeneity. Writing such a protocol is a crucial step as it is intended to reduce biases in data collection. The KLIB cannot provide a complete and comprehensive list of references for meta-analysis, but it can assist the meta-analyst with the elaboration of the protocol in several ways: (i) search and compilation of a list of potential publications that could be included in the meta-analysis and will help with the identification of study inclusion criteria; (ii) identification of different methods used in the studies; (iii) identification of site-related sources of heterogeneity of the results; (iv) identification of studies with extractable data in terms of data quality as the KLIB already presents available metadata on the publications; (v) serving as a filter for study quality as the publications uploaded to the KLIB have already been reviewed by peers of the scientific community.
The KLIB can also assist the meta-analyst in more advanced stages of the meta-analysis, i.e., during the scoping search. The purpose of a scoping search is not to find every study on a particular topic but to identify potential moderators; to decide upon the criteria that would make a study relevant or irrelevant; to identify suitable outcome measures; and to make exclusion decisions based on the title, abstract, or site-related features. The KLIB makes the relevant metadata for publications easily accessible, which will facilitate the scoping search. Additionally, with the network visualization tools, clusters of related publications can be easily identified, which can also speed up the process of searching and decision making. The KLIB has been used in the past to support the scoping searches, as well as for the identification of potential moderators and outcome measures in Betancur-Corredor et al. (2022a, b) and Lang et al. (2023).
2.5.3 Improving process understanding
Another motivation to implement the KLIB was the idea of improving process understanding. Many experimental studies in soil that delve into the impact of various drivers on soil properties or into the interaction between different soil properties lead to correlative connections between the investigated quantities after an adequate statistical analysis of the observations. However, the ultimate goal in most cases is to come up with a better understanding of causalities and their underlying processes. This is particularly true if we are striving for mechanistic, process-based models that are capable of predicting the changes in soil systems in response to external drivers of land use and climatic conditions.
By filtering the available literature according to critical experimental boundary conditions, the KLIB can help to reduce the complexity of the reported results, opening the view to the underlying causal relationships. The example of the impact of reduced tillage on soil carbon given above can also be viewed from this perspective, for instance by identifying one specific driver – such as site conditions or management – to explain the major differences in contradictory study results. This joint driver might help in understanding the underlying processes responsible for the dynamics of soil carbon under different tillage regimes and could then be further tested or integrated into the model description.
2.5.4 Structuring references for individual projects
All three different modes of application are basically useful for any scientific work, be it a classic scientific publication or a PhD thesis. It is an essential part of embedding one's own research into the broader research landscape. Hence, the KLIB can be used for preparing individual publications, project proposals, or theses. It is possible to restrict all visible contents and visualization tools to only those publications that were entered by a single user or a group of authors. In this way, the KLIB can be used as a tool to structure the bibliography of a specific project.
With the KLIB, we offer a tool for structuring soil process research relying on relevant publications including metadata such as site characteristics and experimental boundary conditions. The KLIB can be used freely by anyone interested in soil science, and its value and usefulness will grow with an increasing number of active users and uploaded publications. At the moment, about 600 publications are part of the KLIB. This is sufficient to assess the functionality, but for many questions, the number of publications is still far too small to provide a good overview of the published knowledge. Therefore, all soil scientists are invited to contribute to and ultimately benefit from the KLIB. The visualization modes that make the most of the library are designed to be efficiently applied when the number of publications included in the KLIB grows substantially, which is actually the intended objective.
At the moment, the KLIB is designed for experimental work directed towards a better understanding of soil processes that depend on site characteristics and experimental boundary conditions. The development was motivated by the need for mechanistic soil models (Vogel et al., 2018), such as Expert-N (Engel and Priesack, 1993) or APSIM (Holzworth et al., 2014), that are designed to describe and predict changes in soil functionality based on a sound understanding of the underlying processes. However, these processes and their interactions can hardly be described based on first principles. Instead, their parameterization needs to rely on empirical findings documented in publications on soil process research. The KLIB should be a valuable tool for identifying processes and for coming up with ideas on how to represent these processes in systemic models. In fact, only a smaller proportion of publications within soil science are related to process research. A very substantial part deals with methodological developments. In the future, the KLIB will be further developed for such publications as well in order to also support the literature search for methodological aspects.
Of course, an open-access tool like KLIB raises the question of how to ensure the quality of the data entered by many different users from different fields of soil science. One quality assurance measure is that only registered users can upload publications including metadata. It is completely transparent who has uploaded which publication to the library, which should support the responsible use of the KLIB. The KLIB is maintained by the BonaRes Centre in terms of the quick removal of obvious errors. Further evolution will show whether there is a need for additional functionality to allow all users to mark suspected inconsistent entries as such.
A critical obstacle to uploading publications is the considerable effort required to enter the metadata for each publication, although the user interface supports this process very well, and there is an electronic “cookbook” available on the website. Many entries can be made very quickly via drop-down menus, taking into account different classification schemes and metrics with different physical units. As a side note, this effort is significantly reduced for well-written articles where the information can be found easily.
Despite all this, it seems somewhat anachronistic that, in times of rapid development of artificial intelligence, we are still entering such metadata via the computer keyboard. We are currently testing advanced text-mining tools, as done recently by Blanchy et al. (2023), to minimize the manual data entry. This can work quite well if the authors stick to existing classification systems and nomenclature. The practice has shown, however, that there is an enormous diversity in terminology and that it will take some development before self-learning systems can digest this diversity in such a way that the results meet the requirements of KLIB.
Today, there are a number of web-based tools available which are useful to explore published work in a given research field. Tools such as CONNECTED PAPERS (https://www.connectedpapers.com, last access: 8 October 2023) or Citation Gecko (https://citationgecko.azurewebsites.net/, last access: 8 October 2023) start from a number of seed publications to analyze which other publications have been published on the same subject and how the various publications cite each other. Other tools such as “paper-digest” (https://www.paper-digest.com, last access: 21 July 2023) analyze the contents of individual publications using methods of text mining to summarize their contents. The underlying concepts are developing fast at the moment, and the KLIB will certainly profit from this in terms of automatically extracting information on the spatial coordinates, the soil type, the land use, and other boundary conditions in the future. To the best of our knowledge, the KLIB is the only tool that can be used to relate scientific publications and their major findings directly to site conditions, which is of crucial importance for their interpretation, especially in soil science. Another unique feature of the KLIB is that it can be used by individual users for their own work, but the information they provide is also accessible to other users as well. In this way, we hope that the KLIB will develop into a community venture from which we can all benefit.
The code for the KLIB was generated by the authors and is available from a persistent Git Repository: https://git.ufz.de/Bonares/knowledgelibrary (Vogel et al., 2023).
All mentioned data refer to the content of the literature database KLIB and are publicly available via the corresponding link: https://klibrary.bonares.de (Bonares Knowledge library, 2023).
Conceptualization: HJV, BS, and UW. Data curation: HJV, BBC, SK, BL, ML, ER, BS, and MW. Technical implementation: LF and UW. Writing – original draft: HJV and BBC. Writing – review and editing: all authors.
The contact author has declared that none of the authors has any competing interests.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We would like to thank the many colleagues who gave us a lot of advice on the design of the user interface and functionality during the development of KLIB. This project is funded by the German Federal Ministry of Education and Research (BMBF) in the framework of the funding measure Soil as a Sustainable Resource for the Bioeconomy – BonaRes (project BonaRes (Module B): BonaRes Centre for Soil Research).
This research has been funded by the German Federal Ministry of Education and Research (BMBF) within the framework of the funding measure “Soil as a Sustainable Resource for the Bioeconomy – BonaRes”, project “BonaRes (Module B): BonaRes Centre for Soil Research”.
The article processing charges for this open-access publication were covered by the Helmholtz Centre for Environmental Research – UFZ.
This paper was edited by Ashish Malik and reviewed by two anonymous referees.
Angular: The web development framework for building the future, https://angular.io/ (last access: 16 March 2023), 2023. a
Balesdent, J., Chenu, C., and Balabane, M.: Relationship of soil organic matter dynamics to physical protection and tillage, Soil Till. Res., 53, 215–230, 2000. a
Betancur-Corredor, B., Lang, B., and Russell, D. J.: Organic nitrogen fertilization benefits selected soil fauna in global agroecosystems, Biol. Fertil. Soils, 59, 1–16, 2022a. a
Betancur-Corredor, B., Lang, B., and Russell, D. J.: Reducing tillage intensity benefits the soil micro-and mesofauna in a global meta-analysis, Eur. J. Soil Sci., 73, e13321, https://doi.org/10.1111/ejss.13321, 2022b. a
Blanchy, G., Albrecht, L., Koestel, J., and Garré, S.: Potential of natural language processing for metadata extraction from environmental scientific publications, SOIL, 9, 155–168, https://doi.org/10.5194/soil-9-155-2023, 2023. a
Bonares Knowledge library: KLIB, https://klibrary.bonares.de, last access: 26 July 2023. a
Cote, I. and Jennions, M.: The Procedure of Meta-analysis in a Nutshell, in: Handbook of meta-analysis in ecology and evolution, Princeton University Press: Princeton, New Jersey, 195–206, ISBN 0691137293, 2013. a
Engel, T. and Priesack, E.: Expert-N-A Building Block System of Nitrogen Models as Resource for Advice, Research, Water Management and Policy, in: Integrated Soil and Sediment Research: A Basis for Proper Protection: Selected Proceedings of the First European Conference on Integrated Research for Soil and Sediment Protection and Remediation (EUROSOL), Springer, 503–507, ISBN 978-94-010-4887-3, 1993. a
Guggenberger, G., Thomas, R. J., and Zech, W.: Soil organic matter within earthworm casts of an anecic-endogeic tropical pasture community, Colombia, Appl. Soil Ecol., 3, 263–274, 1996. a
Haddaway, N., Woodcock, P., Macura, B., and Collins, A.: Making literature reviews more reliable through application of lessons from systematic reviews, Conserv. Biol., 29, 1596–1605, 2015. a
Holzworth, D. P., Huth, N. I., deVoil, P. G., Zurcher, E. J., Herrmann, N. I., McLean, G., Chenu, K., van Oosterom, E. J., Snow, V., Murphy, C., Moore, A. D., Brown, H., Whish, J. P. M., Verrall, S., Fainges, J., Bell, L. W., Peake, A. S., Poulton, P. L., Hochman, Z., Thorburn, P. J., Gaydon, D. S., Dalgliesh, N. P., Rodriguez, D., Cox, H., Chapman, S., Doherty, A., Teixeira, E., Sharp, J., Cichota, R., Vogeler, I., Li, F. Y., Wang, E., Hammer, G. L., Robertson, M. J., Dimes, J. P., Whitbread, A. M., Hunt, J., van Rees, H., McClelland, T., Carberry, P. S., Hargreaves, J. N. G., MacLeod, N., McDonald, C., Harsdorf , J., Wedgwood, S., and Keating, B. A.: APSIM–evolution towards a new generation of agricultural systems simulation, Environ. Model. Softw., 62, 327–350, 2014. a
Jarvis, N., Koestel, J., Messing, I., Moeys, J., and Lindahl, A.: Influence of soil, land use and climatic factors on the hydraulic conductivity of soil, Hydrol. Earth Syst. Sci., 17, 5185–5195, https://doi.org/10.5194/hess-17-5185-2013, 2013. a
Koricheva, J., Gurevitch, J., and Mengersen, K.: Handbook of meta-analysis in ecology and evolution, Princeton University Press, ISBN 0691137293, 2013. a
Lang, B., Betancur-Corredor, B., and Russell, D. J.: Earthworms increase mineral soil nitrogen content–a meta-analysis, Soil Organisms, 95, 1–16, 2023. a
Mikolajewicz, N. and Komarova, S. V.: Meta-analytic methodology for basic research: a practical guide, Front. Physiol., 10, 203, 2019. a
Miltner, A., Bombach, P., Schmidt-Brücken, B., and Kästner, M.: SOM genesis: microbial biomass as a significant source, Biogeochemistry, 111, 41–55, 2012. a
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., McGuinness, L. A., Stewart, L. A., Thomas, J., Tricco, A. C., Welch, V. A., Whiting, P., and Moher, D.: The PRISMA 2020 statement: an updated guideline for reporting systematic reviews, Int. J. Surg., 88, 105906, https://doi.org/10.1136/bmj.n71, 2021. a
Schmidt, M. W., Torn, M. S., Abiven, S., Dittmar, T., Guggenberger, G., Janssens, I. A., Kleber, M., Kögel-Knabner, I., Lehmann, J., Manning, D. A., Nannipieri, P., Rasse, D. P., Weiner, S., and Trumbore, S. E.: Persistence of soil organic matter as an ecosystem property, Nature, 478, 49–56, 2011. a
Simpson, Z. P., Jordahl, J., Leptin, A., Miguez, F. E., Niemi, J., Schulte, L. A., Thompson, M. L., Villarino, S. H., and McDaniel, M. D.: No-tillage does not on average reduce soil carbon storage compared to conventional tillage. Comment on “Declines in soil carbon storage under no tillage can be alleviated in the long run” by Cai et al., Geoderma, 430, 116307, https://doi.org/10.1016/j.geoderma.2022.116307, 2023. a
Subirats-Coll, I., Kolshus, K., Turbati, A., Stellato, A., Mietzsch, E., Martini, D., and Zeng, M.: AGROVOC: The linked data concept hub for food and agriculture, Comput. Electron. Agricult., 196, 105965, https://doi.org/10.1016/j.compag.2020.105965, 2022. a
Vogel, H.-J., Bartke, S., Daedlow, K., Helming, K., Kögel-Knabner, I., Lang, B., Rabot, E., Russell, D., Stößel, B., Weller, U., Wiesmeier, M., and Wollschläger, U.: A systemic approach for modeling soil functions, SOIL, 4, 83–92, https://doi.org/10.5194/soil-4-83-2018, 2018. a
Vogel, H.-J., Betancur-Corredor, B., Franke, L., König, S., Lang, B., Lucas, M., Rabot, E., Stößel, B., Weller, U., Wiesmeier, M., and Wollschläger, U.: Code for the KLIB, GitLab [code], https://git.ufz.de/Bonares/knowledgelibrary, last access: 12 October 2023. a
Wachendorf, C., Potthoff, M., Ludwig, B., and Joergensen, R. G.: Effects of addition of maize litter and earthworms on C mineralization and aggregate formation in single and mixed soils differing in soil organic carbon and clay content, Pedobiologia, 57, 161–169, 2014. a