C02: Trees in the rural-urban transitional space of Bengaluru: distribution, functions, significance and relevance.

The focus of C02 in the Phase II is two-fold and grounded in the findings in the Phase I: the subject matter focus is on trees, their pattern and functions along the rural-urban gradient; the methodological focus is on modelling green crown volume of trees and on improving the remote-sensing based classification of “urban”.

We will continue recording fully mapped field sample plots and re-measure a subset of plots for change analyses. We will develop approaches to model tree green crown volume based on terrestrial laser scans, 3D surface models from WorldView3 images, and field observations: we hypothesise that this novel variable explains habitat quality for birds and insects better than the common 2D projection of crown cover. Here, we will collaborate with B01. In collaboration with the Indian counterpart project I-C02, relationships between spectral reflectance, air pollution and tree vitality shall be analysed. In collaboration with FOR2432 projects, relationships between tree variables and biodiversity and socio-economic variables will be evaluated.

We will continue acquiring and analysing WorldView3 imagery both for our tree-focussed research and for upscaling of FOR2432 results in general, where “upscaling” refers to the research transects and to the entire Bengaluru Metropolitan Area from Landsat-8 and Sentinel-2 imagery. We will adjust machine-learning approaches (deep learning) to the automated classification of trees and tree species (groups), and for mapping impervious surfaces. The latter will be the basis for devising a transparent classification and assessment scheme for „urbanness“ as a function of percent built-up.

In line with the FOR2432 data management strategy, C02 will upgrade and maintain the already set up WebGIS that comprises geodata of all FOR2432 projects.

Prof. Dr. C. Kleinn
Dr. N. Nölke
Waldinventur und Fernerkundung
Universität Göttingen

Zihui Zhu
Doctoral Researcher

Indian partner project:
Spatio-temporal land use patterns and the relationship between green areas and biophysical and socio-economic features
B.N. Diwakara, Institute of Wood Science and Technology, Bangalore
V.P. Tewari, Himalayan Forest Research Institute, Shimla
R.R. Nidamanuri, Indian Institute of Space Science and Technology, Trivandrum

Spatio-temporal land use patterns in the rural-urban interface

Green spaces play an important role in urban and peri-urban areas and are the major defining land cover in rural areas. Depending on their relative position along the rural-urban gradient and the surrounding built-up areas, they serve various functions and it can be expected that social-economic variables of particular quarters exhibit a particular relationship to green spaces spatial pattern. Project C02 researches these relationships. As a pre-condition to the quantitative analyses, rural-urban indices are analysed and tested for their suitability to integrate patterns of green spaces as indicators. Remote sensing imagery at different spatial resolutions is being used to map and analyse green spaces for greater Banagalore (Landsat and RapidEye) and to determine biophysical characteristics, uses and functions of green spaces (WorldView 3, in a selected transect). The remote sensing image pre-processing and classification as well as the corresponding ground truthing is implemented in close collaboration with the Indian project collaborators; and social data will be sought from official sources including the Bangalore Development Authority. An image segmentation of greater Banaglore will be carried out into areas of similar green spaces patterns as a basis for establishing relationships between biophysical and social variables. A Landsat based time series of land use will be built in order to test the hypothesis that the changes in green spaces configuration follow a similar pattern both over space and over time. Options will be tested to make the time series more accurate and speficic by sub-pixel analyses for green spaces using up-to-date Landsat imagery and high resolution imagery as a basis for spectral unmixing. In addition, as a prototype for the entire research group, a WebGIS will be established that accommodates all spatial data of C02 and serves as a visualization tool which may support identification of joint thematic interests and therefore foster collaboration between FOR projects.