Comment analyser les données pour identifier les pratiques d’utilisation durable des terres ?

Though this question depends on your research question, in my experience as a researcher in Climate-Smart Agriculture (CSA) interventions (an approach that encompasses many sustainable practices including sustainable land use practices e.g., Agroforestry), key data needed for analyzing sustainable land use practices are, soil data (soil ph, soil organic carbon, etc) water data (water availability, ground water quality, etc) and biodiversity (diverse organisms interaction). These data can obtained through field observations and use of satellite technologies (i.e., GIS & RS).

Depending on the specific research question, comprehensive data on various aspects of land use is essential, including land cover, land use intensity, land degradation, carbon stocks, water quality, and social and economic indicators. Possible data sources encompass remote sensing data obtained from satellites or aircraft, GIS data stored in geographic information systems, and field data collected through direct observation or measurement. Remote sensing data is valuable for information on land cover types and vegetation indices, GIS data provides insights into land use patterns and suitability, and field data offers information on land use impacts, ecosystem functions, biodiversity status, and human well-being.

The SBTI Forest, Land, and Agriculture (FLAG) guideline plays an important role in addressing climate change by reducing emissions and sequestering carbon. Not only is FLAG one of the sectors at highest risk from the impact of climate change, but it is also a significant source of emissions. It represents nearly a quarter of global greenhouse gas (GHG) emissions - the largest emitting sector after energy. It encompasses a range of activities including forest management, agriculture, and land use change, all of which have significant impacts on global GHG emissions. Companies operating in the FLAG sector should therefore measure and manage their emissions to effectively target climate change, to the transition to a low-carbon economy.

Ensuring data quality is crucial; it involves checking and correcting the data for errors, outliers, missing values, duplicates, or inconsistencies to guarantee accuracy and reliability. Data harmonization is also necessary to make the data compatible and comparable across different sources and scales by converting units, resolutions, projections, classifications, or time periods to a common standard or reference. Lastly, data integration involves combining data from diverse sources and formats into a unified dataset or database through processes like joining, merging, appending, or overlaying the data based on common attributes or spatial locations.

For FLAG screening as first step the goal is to review the operational and value chain model of the company to determine FLAG & Non-FLAG boundaries: 1) FLAG & Energy/Industry (Non-FLAG) inventory boundaries determined 2)SBTi target setting strategy confirmed across FLAG & Energy/Industry (original near-term target criteria) standards After that execute the relevant FLAG specific calculations to fulfill land sector specific requirements: 1) Relevant FLAG (Land Use Change, Land Management CO2, Land Management Non-CO2) GHG calculations conducted 2)Specific FLAG related methodologies fully documented with full transparency

"All models are wrong, but some are useful" ... based on this times tested and well applied principle, a predictive model should be developed using geo-spatial data and integrating field data into it in an innovative and out-of-the-box approach. Followed by accuracy assessment using ground-truthing. I have done that in several projects. Believe me, there is no alternative to get your feet dirty; the hard and bold way.

This may include applying various statistical, spatial, or modelling techniques to explore, describe, explain, or predict the relationships between land use and sustainability indicators. Descriptive analysis is used to summarize and present the data through tables, charts, maps, or graphs, providing an overview of the data distribution, trends, patterns, or variations. Inferential analysis tests and estimates the data using hypotheses, significance tests, confidence intervals, or correlation coefficients to offer evidence of causality, association, or difference. Lastly, predictive analysis forecasts and projects the data using regression, classification, or simulation models to provide scenarios of outcomes, impacts, or risks.

Aggregating the bottom-up FLAG & Non-FLAG calculations to generate two consistent and coherent inventories with 2 main outcomes: 1) FLAG & Energy/Industry inventories created 2) SBTi FLAG specific data points generated to support validation

Focus is disaggregating FLAG Emissions: 1) Screening - Emission Factor (EF) analysis to determine FLAG relevance. Example: Milk 2) Categorisation - Classifying specific FLAG & Non-FLAG sub-categories. Example: Land Use Change - N2O, from biomass - > CO2, from Biomass 3) Consolidation - LUC, Biogenic - > LUC, Non-Biogenic

The fourth step in analyzing data for sustainable land use practices is to visualize and communicate the data results and insights effectively. This involves creating and designing engaging visualizations using tools such as Excel, Tableau, Power BI, QGIS, R, or Python. Adhering to data visualization principles, such as selecting the right chart type, employing appropriate colors and scales, labeling and annotating the data, and providing context and explanation, is crucial to ensuring clarity, accuracy, and relevance in communicating the findings.

This involves following and applying ethical principles and standards governing the collection, processing, analysis, visualization, and communication of data. Addressing data ethics issues and questions, such as data privacy, ownership, quality, and impact, is essential. This includes obtaining consent or anonymizing data for privacy, citing sources or attributing credits for ownership, verifying data or reporting errors for quality, and evaluating outcomes or mitigating risks for impact. These measures protect personal and sensitive data, recognize the rights of data owners or providers, ensure the validity and reliability of the data, and assess potential positive or negative effects.

When I read these responses I see echoes of AI. At UCB and UCD we learned about sustainability before a lot of the technology and climate change waves ever broke. So understanding the Earth, it's landforms, biota, and human impacts underlies any technological or programmatic approach.. Diverse international and applied studies are key elements, and up-front ethics, not just buzzwords apply. The recent conference spoke to fossil fuels but overlooked land use and cultural transformation.

FLAG targets are additional to Energy and Industry targets Companies from the following SBTi-designated sectors: Forest and paper products (forestry, timber, pulp and paper); food production (agricultural production); food production (animal source); food and beverage processing; food and staples retailing; and tobacco. Companies in any other sector with FLAG-related emissions that total. While several reporting guidelines such as the GRI and ISSB are in place to create consistency in emissions reporting, there is still no universal format for measuring and reporting sustainability performance more than 20% of overall emissions across all scopes. 20% threshold should be accounted for as gross emissions, not net (gross minus removals).

The old carpenter's saying, "measure twice, cut once," really applies here. If you have not been very careful and clear in how you define "sustainable" land use practices, then your methodology, data collection, and analysis will be problematic. Make sure you are setting out a clear conceptual definition that you can understand, define, and defend before you begin to look for and analyze data. Otherwise you might find yourself looking for a question to fit the data rather than seeing how data helps to answer (or not) your question.