A systematic review on remote sensing of dryland ecological integrity: improvement in the spatiotemporal monitoring of vegetation is required

Highlights: What are the main findings? Remote sensing of ecological integrity in drylands relied heavily on indicators related to vegetation. Novel approaches that address key limitations of vegetation remote sensing in drylands (e.g., the background soil effect, spectral similarity between non-photosynthetic vegetation and bare soil) are increasingly used in monitoring applications. What are the implications of the main findings? The review identified key knowledge gaps, particularly in the relationships between novel vegetation monitoring approaches and indicators of ecosystem quality, as well as in the effects of management practices. Novel approaches to dryland vegetation monitoring also need further development, in terms of benchmarking and field validation, to become reliably operational. Remote sensing approaches to monitoring dryland ecosystem states and trends have been dominated by the binary distinction between degraded/non-degraded areas, leading to inconsistent results. We propose a different conceptual framework that better reflects the states and pressures of these ecosystems—ecological integrity—that is, the maintenance of ecosystem composition and its capacity to contribute to human needs and adapt to change. We systematically reviewed earth observation techniques for characterizing ecological integrity in trusted databases together with studies identified through expert-guided search. A total of 137 papers were included, and their metadata (i.e., location, year) and data (i.e., aspect of ecological integrity assessed, techniques employed) were analyzed. The results show that remote sensing ecological integrity is becoming an increasingly researched topic, especially in countries with extensive drylands. Vegetation was the most frequently monitored attribute and was often employed as an indicator of other attributes (i.e., soil and water quality) and as a key feature in approaches that aimed for a comprehensive ecosystem assessment. However, most of the literature employed the normalized difference vegetation index (NDVI) as a descriptor of vegetation characteristics (i.e., health, structure, cover), which has been shown not to be a good indicator of the litter/senescent vegetation components that tend to frequently dominate drylands. Methods to overcome this weakness have been identified, although more research is needed to demonstrate their application in ecological integrity monitoring. Specifically, knowledge gaps in the relationship between vegetation cover fractions (i.e., green, non-green, and bare soil), descriptors of ecosystem quality (e.g., soil condition or vegetation structure complexity), and management (i.e., how human intervention affects ecosystem quality) should be addressed. Notable potential has been identified in time series analysis as a means of operationalising remotely sensed vegetation fractional cover. Nevertheless, limitations in benchmarking must also be tackled for effective ecological integrity monitoring.