Theories about critical transitions and threshold-dependent changes (TDC) in natural systems have been largely developed in fluid systems using bifurcation theory. Application of these theories to terrestrial systems has not been as convincing, possibly because they differ from fluid systems in terms of their levels of spatial interconnectivity. Here, we use new theory to understand the mechanisms of decline in already 'degraded' terrestrial systems. Terrestrial systems consist of varying numbers of discrete spatial units (modules) with low interconnectivity between them, whereas fluid systems approach a single spatial unit with near ubiquitous interconnectivity. Furthermore, bifurcation theory applies optimally to self-organised and complex system states rather than simplified, homogenised or 'degraded' states. We hypothesise that, due to lower spatial interconnectivity (i.e., more discrete spatial units), terrestrial systems are more likely to undergo Turing bifurcations (localised state changes which arise due to differential diffusion, giving rise to spatial patterns) rather than the conventional 'saddle-node'/fold bifurcations (a local bifurcation in which two equilibria of a dynamical system collide). Degraded terrestrial systems, such as agroecosystems, may therefore be quasi-stable states where further degradation might be characterised by spatial reorganisation predicted by Turing bifurcations rather than the abrupt decline expected following 'saddle-node'/fold bifurcations. This PhD focusses on agroecosystems to provide new evidence to understand the mechanisms that maintain system stability, lead to further decline or recovery, and control the rates of change in overall system state. We will capture near real-time imaging over on-going resilience plot experiments. We will combine these primary data with existing spatiotemporal datasets from the North Wyke Farm Platform (NWFP) already captured at the sub-field, field, farm and landscape scale, as national/continental satellite data. Combined, these datasets allow for the detection of the localised abrupt changes within discrete spatial units (evidence of spatial reorganisation) that are anticipated under Turing bifurcations.