Revisiting the Landscape Mosaic model

Abstract

The landscape mosaic model quantifies and maps the spatial juxtaposition of different land uses. It provides a landscape perspective of anthropic threats posed by agriculture and urban development, and the spatial-temporal shifting of the landscape mosaic indicates landscapes where anthropic intensity has changed. We use the U.S. Geological Survey provided National Land Cover Database (NLCD) for the years 2001 and 2021 to derive the landscape mosaic at five analysis scales. To improve earlier implementations of the model, we introduce the heatmap, a flexible scheme providing more thematic reporting opportunities and allowing for better quantitative summary reporting across analysis scales as well as for temporal trends. The results are exemplified at regional scale for the Atlanta metropolitan area. We use the improved model to investigate the land cover context over time and at different analysis scales and show how custom color tables detail different thematic features of the landscape mosaic, including the degree and change of anthropic intensity. We conclude with a discussion of potential applications in ecology, landscape planning, and restoration to illustrate the benefits of the revised landscape mosaic model. All assessment tools are now available in open-source software packages.

Fig 1

Legend for the original 19-class LM map (a), and the new enhanced 103-class LM map (b). Numbers in (b) are class codes.

Fig 2. Explanation of how to read the Landscape Mosaic pixel triplet values, denoting their relative contribution to blue, green, and red (here agriculture, natural, and developed) at a given location and for the local moving window area in use.

Fig 3. LM around Atlanta for the year 2021 using a moving window size of 13 pixels.

Showing the 103-class LM map and applying different color tables to highlight dominant background land cover (a), degree of diversity (b), or anthropic intensity (c).

Fig 4

LM 103-class mapping to the degree of anthropic intensity (a); associated 103-class pixel color scheme (b); and respective anthropic intensity value (c).

Fig 5

Anthropic intensity for Atlanta 2001 and 2021 (a); histogram of change pixels and anthropic intensity change color range legend (b); locations and degree of anthropic intensity change (c). Results based on 13-pixel window size (~15 ha).

Fig 6. Temporal change (2001–2021) of LM around Atlanta using a moving window size of 13 pixels.

LM heatmap for 2001 (a), LM heatmap for 2021 (b), and the delta heatmap showing change in LM-class percentages between the two LM heatmaps with the maximum increase and decrease highlighted in black circles (c).

Fig 7

LM for the year 2021 around Atlanta: comparison of LM-classes across five analysis scales of moving window size (a); summary proportions per scale and difference in LM-Background and refined Mixed (b); delta heatmap between smallest and largest scale with the maximum increase and decrease highlighted in black circles (c).