Cerro Seco - Digital Soil Mapping

In this repository you will find:

• Field and laboratory data

• Spatial data

• R code

• Simplified reports

• Visualizations

The data comes from a citizen-based soil survey in the area known as Cerro Seco (southern Bogotá). It was collected between 06.2020 and 07.2021. Funds for this project were given by the True Cost Initiative to Corporación Geoambiental Terrae (NGO), with whom I collaborate.

This study aimed to create a preliminary map of ancient soil (paleosol) sequences distribution in the area, using machine learning classification. Two models were selected due to time and data availability: K-nearest neighbors and Random Forest. The map was integrated with hydrological models of typical soil profiles and of the underlying rock, to make a first estimate of water recharge and overflow/interflow in the area. Information was submitted to the Secretaría Distrital de Planeación and the Secretaría Distrital de Ambiente to support the creation and management of a new conservation area in Cerro Seco.

Study locations · Ubicaciones de estudio

Model profiles of paleosol/hydropedologic sequences Modelos de perfiles de suelo

Random Forest Importance of remote sensing-derived indices Importancia de índices

Preliminary sequence distribution models Modelos preliminares de distribución de secuencias