02/06/2019
International Spring University Bilbao, Spain
by James Wafula
From 19th to 25th May 2019 I attended a workshop at the International Spring University (ISU) in Bilbao, Spain. The workshop was designed to introduce the participants to:
• ARIES, k.explorer and k.Lab
• The ARIES environment and available models
• The technical principles ARIES is based on
• The results and knowledge ARIES can provide
• How ARIES addresses the current practice and decision-making
• Examples of ARIES applications
• How I can use ARIES to solve my current (research) questions in the River Nzoia Basin.
I am currently carrying out a research in Kitale located Western Kenya. My research area covers 2 wards in Trans Nzoia County which is part of the larger River Nzoia Basin stretching from the slopes of Mt. Elgon to the shores of Lake Victoria, an area of approximately 12,000 km2. As a Climate Innovation fellow with AfriClp, I am involved in the production and promotion of improved organic fertilizers from livestock manure, domestic wastes and selected leguminous plant species. The ultimate goal of my project is to reduce the consumption and need for chemical fertilizers which are recognized as one of the causes of extensive nutrient loading of the river and lake basin in addition to contributing towards ground water contamination.
The ecosystem services I am primarily interested in are:
1. Water quality regulation.
2. Sedimentation.
ARIES stands for Artificial Intelligence for Ecosystem Services and it is the flagship tool that uses the modelling environment k.Lab (which stands for knowledge laboratory). It makes use of multiple modeling techniques from multiple scientific fields such as spatial mapping, hydrology, plant physiology, and soil sciences to name but a few. By making use of artificial intelligence applied to semantic modelling, machine reasoning, and machine learning, ARIES is able to connect the data and models from multiple sources to provide a spatially explicit ecosystem services output model which can be mapped out. Ecosystem services output models may involve natural capital models (such as arable land, carbon dioxide absorption, erosion control, minerals, water, and waste assimilation), natural processes (such as sediment transport, nutrient cycles, and floods) and demand models on the human beneficiaries’side (farmers and special interest groups). These models may provide the knowledge base to further value and manage ecosystems around potential future scenarios.
ARIES can be used by researchers and practitioners to quantify, map, and evaluate ecosystem services and the beneficiaries of these services in case studies covering areas such as the Mt. Elgon ecosystem, and the River Nzoia Basin. Some of the case studies have been designed using locally available, high-resolution spatial datasets to populate models that represent a broad range of ecosystem services in a variety of ecological and socio-economic settings. A relevant case study might include a modeling approach that uses information on the location of crops, nitrogen fertilizer and pesticide applications, soil characteristics, and depth to the ground water table to predict nitrate and pesticide concentrations in wells.
The user-friendly and explorer side of ARIES is the k.explorer. This explorer allows first contact users to query and view live requests for existing models within the system through an easy searchable space bar and drag and drop utilities. However, users are also able to modify existing models, fit them to specific purposes and tailor them by using the modeler side of ARIES which is the k.LAB software modelling environment. This software consists of a set of tools that allows the user to develop models using the k.IM semantically driven language. The k.LAB software can therefore be described as a general model development environment that provides access to a vast network of servers and modeling engines and to specialized tools to create and test one’s own tailored projects.
The ISU training was essentially meant to introduce me to the ARIES k.explorer and to show me how to begin using its set of tools. To begin developing models, one has to be familiar with the k.IM language and the basics of semantically integrated modeling.
Semantic Annotation, also known as semantic tagging or semantic enrichment is the process of attaching additional information to various concepts (e.g. people, places, organizations, etc) in a given text, map, video, or any other content. Unlike conventional text annotations which are for the readers’ reference, semantic annotations are used by machines (computers). When a document (or another piece of content e.g. a GIS map) is semantically tagged, it becomes a source of information that is easy to interpret, combine, interoperate and reuse by our computers. Semantic Annotation makes it easy to:
• Find relevant information in heaps of documents with the help of machines doing the leg work.
• Extract knowledge from disparate sources.
• Provide personalized content, based on machine understandable context.
• Automatically interconnect data.
This then brings us to the concept of the Semantic Web, at the heart of which is making data and models interoperable (through the FAIR- findability, accessibility, interoperability, and reusability- principles). This concept allows the research community to:
1. Put data and models online, where they can be found by people and computers (through search)
2. Label the data and models consistently, so people and computers know what they are, without having to make guesses
3. Develop and use apps that can assemble data and models, in a system of ontologies , to fit a user’s needs for different places, times, and scales.