Discrete and continuous approaches to characterizing subannual urbanization dynamics from multi-scene, multi-decadal Landsat imagery
Abstract
The Greater Houston area has experienced extraordinary growth over the past two decades, characterized by steep increases in population as well as land cover conversion on an historically unprecedented scale. Owing to the imperative to better understand the interactions of global climate change and local land cover change, researchers have been tasked with assessing how the conversion of natural and semi-natural cover types to impervious surfaces can exacerbate the impact of natural disasters in large urban areas, like that of Hurricane Harvey on Houston in 2017. In this vein, in this study we leverage over twenty years of Landsat satellite observation to synoptically characterize the spatio-temporal dynamics of land cover change in the Greater Houston area on an annual and sub-annual basis from 1997-2018. We adopt a unique automated classification procedure that combines adaptive signature generalization, multi-scene compositing, ensemble classification using all acceptable imagery in a given year, as well as spatio-temporal stabilization for consistency across sensors and scenes. In so doing, we focus on two alternative automated approaches to remotely-sensed, urbanization time series: temporal trajectories of discrete land cover types (classification) versus that of continuous subpixel impervious fractions (regression). Land cover products are verified with NLCD products and validated using independent high-resolution aerial imagery, showing per-class accuracies between 74-94%. Validated maps corroborate what census and economic data likewise indicate: Greater Houston grew by 31% over the past 20 years, 14% of which occurred in the FEMA 100-year floodplain. Results demonstrate the added value of multi-decadal, temporally-dense time series for characterizing higher-order spatio-temporal dynamics of land cover, including periodicity, abrupt transitions, and time lags emerging from underlying demographic and socio-economic trends.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.B13B..05H
- Keywords:
-
- 0439 Ecosystems;
- structure and dynamics;
- BIOGEOSCIENCESDE: 0480 Remote sensing;
- BIOGEOSCIENCESDE: 1632 Land cover change;
- GLOBAL CHANGEDE: 1807 Climate impacts;
- HYDROLOGY