Triton is one of the few satellites in the Solar System which shows an ongoing geological activity, with plumes and geysers whose origin is still controversial [1]. Its surface is relatively young, as shown by the paucity of craters detected on its surface [2]. Our knowledge of this moon comes from the Voyager 2 mission, which obtained several images, covering about 40% of Triton’s surface [3]. However, few studies were focused principally on surficial geomorphology, and those are mostly limited to the cantaloupe region and surrounding areas. Triton’s crust is composed predominantly by solid nitrogen (N2) but several other ices have been detected [3]. Crater counting has revealed that the surface is very young and likely it went through a resurfacing process in the past. In fact, a very small number of craters has been detected, and these usually exhibit a typical bowl-like shape [2]. Geological features on Triton include regions, called terrains, such as cantaloupe terrains or plains, which show different textures. Usually, plains are categorized within smooth, walled and terraced plains [3]. The latter are the flattest areas on Triton, a characteristic which has been explained by evoking a lava-like or other viscous liquid infill. Their central depressions also present a cluster of irregular pits, which have been interpreted as drainage pits or eruptive vents [3]. These peculiar morphologies seem to indicate the presence of a viscous fluid on the surface in a remote epoch, which may imply potential climatic and atmospheric changes during Triton’s geological history. In this work we analyse an area located at NW of Tuonela Planitia, which shows several depressions rimmed by sharp margins. Two of these depressions are named Kulilu Cavus and Mah Cavus [4]. Cavi are elliptical-shaped depressions, distributed in an ordered trend, which constitute the cantaloupe terrain [3]. Diapirism is the main candidate process to explain these collapsed depressions [3] but other hypotheses, such as cryovolcanism or impact cratering [5], have also been proposed. Methodology A new geological map has been realized. We used Voyager 2 imagery named c1139533 [6] (600 m/px), properly calibrated, filtered and georeferenced using the Integrated Software for Imagers and Spectrometers (ISIS4) [7]. We mapped the different geological units and main features according to differences in surface morphology (fig.1). We also produced a DEM of the study area, using the open-source suite of tools NASA Ames Stereo Pipeline (ASP) [8]. We applied the photoclinometry-based “shape-from-shading” (SfS) tool to produce the DEM. Since SfS needs an input DEM generated preferably with stereo images, and we do not have such data for Triton, we used the methodology proposed by Lesage et al. 2021[9]. We analysed four different cross sections to measure the relative height of Kulilu Cavus, Mah Cavus and two other depressions, as well as their associated terraces (fig.2). Discussion Geologic