Abstract

Oasis ecosystems are highly vulnerable to environmental changes. To determine the state of vegetation in these ecosystems, monitoring systems must be provided with data on cultivated areas. These data can be obtained in part by using satellite observation systems with high and moderate spatial resolution and high temporal repetitiveness; these systems offer a synoptic vision that makes them a particularly appropriate information source for effectively estimating such data. In this study, we describe an approach to monitor the changing dynamics of Tozeur oases in southwestern Tunisia. To this end, we used a time series decomposition method (seasonal and trend decomposition using loess) to extract the trends from a multi-year time series at the scale of a geographical point (250 m × 250 m pixel) across the MOD13Q1 time series (2000–2016) of the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor, at a 250-m spatial resolution time series. These methods were tested with the final aim of setting up an oasis monitoring system based on the analysis of time signatures obtained from MODIS images. The results showed that it was possible to identify the main types of irrigated perimeters present in the Djerid region and retrospectively trace their recent development history.

摘要

绿洲生态系统极易受到环境变化的影响。为了确定这些生态系统中的植被状况，必须向监测系统提供耕地面积数据。这些数据可以部分通过使用具有高、中空间分辨率和高时间重复性的卫星观测系统获得；这些系统提供了一种概括性的视野，使它们成为有效估计此类数据的特别合适的信息来源。在这项研究中，我们描述了一种监测突尼斯西南部托泽尔绿洲动态变化的方法。为此，我们使用时间序列分解方法（使用 loess 进行季节性和趋势分解）从 MOD13Q1 时间序列（2000-2016 ) 的中分辨率成像光谱仪 (MODIS) 传感器，在 250 米空间分辨率时间序列。对这些方法进行了测试，最终目的是建立一个基于对 MODIS 图像时间特征分析的绿洲监测系统。结果表明，可以确定杰里德地区存在的主要灌溉周长类型，并追溯其最近的发展历史。测试这些方法的最终目的是建立一个基于对从 MODIS 图像获得的时间特征的分析的绿洲监测系统。结果表明，可以确定杰里德地区存在的主要灌溉周长类型，并追溯其最近的发展历史。对这些方法进行了测试，最终目的是建立一个基于对 MODIS 图像时间特征分析的绿洲监测系统。结果表明，可以确定杰里德地区存在的主要灌溉周长类型，并追溯其最近的发展历史。