Application of Remote Sensing for Ex Ante Decision Support and Evaluating Impact

Abstract

Deforestation is an environmental stressor that dramatically threatens biodiversity while having adverse, deep-rooted, socioeconomic impacts. Interventions to address deforestation require considerable long-term investments; therefore, assessing their impact is crucial. The UN Convention on Biological Diversity encourages establishing protected areas (PAs) as a strategy to conserve biodiversity, secure ecosystem services, support local livelihoods, and reduce climate change impacts. The Global Environment Facility (GEF) has supported the conservation and management of PAs, yet site accessibility and data availability present challenges in assessing PAs using traditional evaluation methods. We present a novel application of satellite Earth observations for evaluating the effectiveness and impact of PAs as a deforestation mitigation strategy. Geospatial analysis and ecological forecasting methodologies provide an efficient way to quantify land cover change, estimate aboveground carbon stock, and evaluate ecosystem services provided by the PAs ex ante. We demonstrate the value of a geospatial approach through the evaluation of GEF-supported PAs in Kenya.

A special delivery by a fork: Where does artificial intelligence come from?

“…Recent examples of neural networks in use in evaluation studies include Chan and Chau (2019), Gao (2021), Thieme et al. (2022), and Zhou et al. (2019), among others.…”

“…Neural networks also power some AI summarization tools that condense a large body of text by recognizing its most important segments (Nallapati etal., 2017). Recent examples of neural networks in use in evaluation studies include Chan and Chau (2019), Gao (2021), Thieme et al (2022), andZhou et al (2019), among others.…”

Evaluating site selection at design in food systems interventions: A formative geospatial approach

“…Geospatial analysis has rarely been used for formative environmental evaluation given the difficulty of seeing stress reduction evidence so soon into implementation. Thieme et al (2020) used geospatial land use change modeling to predict where future deforestation might occur near intervention areas, but this served to show that further conservation investment could be needed rather than as part of a formative evaluation.…”