Abstract:

In this study, we investigated the use of  Landsat 8 and Sentinel-1 time series data for detecting disturbances in Madagascar using machine learning algorithm in cloud based Google Earth Engine (GEE) Platform. The random forest algorithm was used to predict the land cover change from 2000 - 2018. The results showed that Madagascar lost 21.02% of humid forest was lost between 2000 - 2018. This tremendous loss is due to disturbances caused by slash and burn agriculture, Urbanization and logging. Approximate, 1.21Giga tonnes of  CO₂ was released into the atmosphere as a result of tree cover loss in Madagascar. The overall classification accuracy achieved is 90.2%. 

​

Reference:

GOFC-GOLD, 2016, A sourcebook of methods and procedures for monitoring and reporting anthropogenic greenhouse gas emissions and removals associated with deforestation, gains and losses of carbon stocks in forests remaining forests, and forestation. GOFC-GOLD Report version COP22-1, (GOFC-GOLD Land Cover Project Office, Wageningen University, The Netherlands).

Abstract:

The AFOLU Programs works in conjunction with Climate, Community and Biodiversity standards (CCB). The CCB standards can be applied to all types of land management project, including reforestation, afforestation, revegetation, forest restoration, agroforestry, sustainable agriculture, and other land management. The Gold status is awarded to projects that satisfy one of the optional criteria by providing exceptional benefits including explicit design for adaptation to climate change, benefits for globally poorer communities, or conservation of biodiversity at sites of global conservation significance. This study is focused on demonstrating Gold level optional criterion (GL1) for demonstrating climate change adaptation benefits by project developers using MaxEnt approach. The example here include habitat suitability modelling of White Headed vulture in Ethiopia. The White headed Vulture is Critically Endangered V3.1 (IUCN, 2012) and hence qualifies for this study. In this study, we observed that White headed Vulture would loose about 20-25% of suitable habitats between current (2000) and future (2070) due to climate change. The overall accuracy achieved was fair AUC = 0.791. 

 

Photo credits:Evan Buechley.

Abstract

Chromolaena odorata is a very widely distributed tropical shrub and considered as an highly invasive weed of field crops and natural environments in its introduced range. It continues to spread due to its effective short- and long-distance dispersal. It can form pure stands where established, often in disturbed areas, grasslands, fallow areas and forestry plantations, and is highly competitive. It has been reported to be the most problematic invasive species within protected rain-forests in Africa. In Western Africa it is well known for preventing the regeneration of tree species in areas of shifting cultivation. It affects species diversity in southern Africa. The plant's flammability greatly affects the forest edges. In Sri Lanka it is considered as a major weed in disturbed areas and coconut plantations. 

Chromolaena odorata is still expanding its range, and is considered one of the world’s worst weeds. It is viewed as a major environmental weed, but is appreciated by some agriculturalists as it shortens fallow time in shifting cultivation. The Maximum Entropy (MaxEnt) approach is used for species distribution modelling for present and future (MIROC5) conditions under RCP (Representative concentration pathways) 8.5 for 2050 (average for 2041-2060) and 2070 (average for 2061-2080). The ​Results showed that good level of predictive performance as indicated by the mean AUC value (10 replicated runs) is 0.895 ± 0.0044.