Dr. Cho-Yin Wu 吳卓穎 - Prediction of soil organic carbon contents in Taiwan agricultural soils by Vis-NIR spectrometry 以可見-近紅外光譜推估台灣農田土壤有機碳含量

Regarding the goal of net-zero of carbon emission, soil is the largest natural sink of carbon in the nature. Thus, a rapid and efficient measurement of soil organic carbon (SOC) is crucial to estimate soil caron sink. The conventional methods of measuring SOC are time-consumption and would generate waste liquids or solids such as wet-oxidation, dry combustion, and TOC analyzers. On the contrary, Vis-NIR is a rapid and non-destructive technique for measuring SOC, which is easy to operate and environmentally friendly for offering effective data. Moreover, with a huge amount of soil samples, the approach of Vis-NIR not only is a global and modern trend for SOC determination, but a powerful potential strategy for evaluating soil carbon stocks. This study measured the SOC content of soil samples from different type of soils throughout Taiwan with the Walkley-Black method and a TOC analyzer as well as their Vis-NIR spectra have been obtained with the well-constructed standard operating procedure. Furthermore, these datasets were successfully applied in the Cubist decision tree for building the machine-learning-base SOC prediction model, which can rapidly and precisely estimate the SOC content in soils from Taiwan. 

在2050淨零碳排目標下，土壤被視為最大的自然碳匯，因此快速而有效率地量測土壤有機碳是估算土壤碳匯的成敗關鍵。以往傳統的濕式氧化法測定土壤有機碳，耗時，且會衍生廢液或廢棄物。可見-近紅外光譜法(Vis-NIR)是一個快速且非破壞性的測定土壤有機碳技術，可提供簡便與環保的檢測結果。在樣品數量龐大下，使用Vis-NIR測定土壤屬性不僅是國際趨勢，更是個具有潛力的評估碳匯工具。本研究利用重鉻酸鉀法及TOC儀分析之土壤有機碳數據，結合Vis-NIR光譜資料，以Cubist迴歸樹建構透過Vis-NIR光譜快速推估臺灣土壤有機碳含量的機器學習模型。

Mr. Ernest Habanabakize - Data driven Insights for Sustainable Agriculture: Regenerative Agriculture (RA) in Rwanda

Preliminary field evidence and literature suggest that regenerative agriculture (RA) has a wide range of benefits to people and the planet. But high-quality data and evidence are needed to inform policy and investment frameworks and galvanize support for RA. While many definitions of RA exist in the literature, our approach is based on the 9 principles (practices) of RA established by Andhra Pradesh Community Managed Natural Farming Programme (APCNF) which represents over one million farmers who shifted away from chemical-based farming practices and transitioned to RA over the last decade. Future Earth is collaborating with multiple partners in Rwanda, Canada, India, and the US on the Data-Driven Insights for Sustainable Agriculture (DISA) project. In this event we will discuss how Future Earth Canada & SDA, through DISA, is working towards building innovative, evidence-based decision-making frameworks that leverage digital technologies, governance, and local knowledge to support transitions away from resource-intensive farming models towards knowledge-intensive, nature-positive, net-zero models that drive economies and the role of partnerships to achieve these objectives.