GEOG Seminar 10/2: Alexandra (Sasha) Tyukavina, "General Principles of Land Cover Map Accuracy"

Join us Thursday, Oct. 2 at 3:30 p.m. at River Road or via Zoom for a seminar with Associate Research Professor Alexandra (Sasha) Tyukavina.

Associate Research Professor Alexandra (Sasha) Tyukavina will present updated good practice guidelines and a unified set of estimators for land cover map accuracy assessment and area estimation.

Abstract

Over the past two decades, the importance of assessing land cover map accuracy using an independent reference sample has been emphasized in several key methodological publications, such as Olofsson et al. (2014). As a result, an increasing number of recently published maps are accompanied by some form of accuracy assessment (Stehman and Foody, 2019). However, until recently, there has been no comprehensive synthesis of the latest advancements in land cover map accuracy assessment methodologies and standards in reporting accuracy.

In this seminar, Dr. Alexandra (Sasha) Tyukavina, Associate Research Professor at the GLAD Laboratory in the Department of Geographical Sciences, will present her recent work on synthesizing current best practices in this area. She will provide an overview of the newly published CEOS LPV Global Land Cover Validation Guidelines, for which she served as an editor, and discuss her recent review paper on global sampling methods. Her presentation will highlight a unified set of equations for estimating both map accuracy and land cover class area, offering valuable insights for graduate students and experienced researchers alike.
 

Speaker Bio

Dr. Alexandra (Sasha) Tyukavina is an Associate Research Professor at the Department of Geographical Sciences. She is originally from Russia, and received MSc in Cartography in 2011 from Lomonosov Moscow State University. Sasha's Master's thesis was focused on mapping the northernmost forests in the world (Ary-Mas forest island). She received her Ph.D. in 2015 from the UMD Department of Geographical Sciences under the supervision of Prof. Matthew Hansen. Her dissertation research was investigating forest disturbance dynamics in the humid tropics using optical and Lidar remotely sensed data sets.

During her postdoctoral research, Dr. Tyukavina was working primarily on sample-based estimation of forest loss drivers in the tropics. Since then her focus shifted to characterizing global drivers of forest loss and global sampling methods for accuracy assessment and area estimation. Sasha has published the first global map of forest loss due to fire in 30m resolution. She has lead a NASA-funded research project quantifying the global hotspots of forest loss and their drivers using high resolution data. Sasha's contributions to the field have been recognized with a 2022 Leading Women in Machine Learning for Earth Observation (ML4EO) Award from the Radiant Earth Foundation.

Since 2021 Dr. Tyukavina has been serving as a co-lead of the Land Cover focus area within the Committee on Earth Observation Satellites (CEOS) Land Product Validation (LPV) subgroup. In this role Sasha has initiated the update of the global land cover validation guidelines, which has been last performed in 2006.
 

Location: For Zoom details, please visit the GEOG Department Calendar.