In the last decade or so, digital twins have emerged as a disruptive technology opening doors for digital innovation and modernization. A digital twin is defined as an accurate virtual model reflecting the state of a physical asset. With their futuristic applications, digital twins have already revolutionized several industrial sectors including product life cycle management, smart cities, healthcare and aerospace. More recently, researchers have started creating digital twins for natural environments. Virtual models are developed to understand, manage and protect these ecosystems which have huge potential in resisting their global decline in the face of the environmental crisis. Therefore, this article surveys the state-of-the-art research that is available in the area of digital twins for natural environments from a networking perspective. A comprehensive background on the topic is presented along with a systematic literature survey. Based on the survey, a set of open issues are identified and discussed for enhancing the future work in the field.