Remote sensing land use analyses are an indispensable method for geography, environmental science, and earth system science. Against the backdrop of the increasing climate and environmental crisis, these disciplines and their methods are gaining enormous importance in environmental policy. However, they often have a weak epistemological basis. While they are frequently viewed as technical and objective results, they are in fact the result of socio-technical practices in which the values, decisions, and practices of the creator are inextricably interwoven. This misinterpretation can lead environmental managers to place too much trust in remote sensing’s claims about the world. In two recent articles, I have both demonstrated and theoretically framed the importance of practices and values in land use mapping and shown that remote sensing often fails in three respects. First, it overlooks the importance of epistemological foundations; second, it overlooks the importance of reflexive research practice; and third, it lacks a critical technoethics. An important but still missing step in this argument is to use a concrete research example to reveal what these deficiencies look like and to make practical suggestions as to how they can be remedied. In this article, I thus deconstruct and critique one of my own published classification approaches and outline a better set of research practices for future remote sensing.
Andreas Christian Braun (Wed,) studied this question.