The latest in our series of blogs to accompany the Special Feature, Toward prediction in the restoration of biodiveristy, is written by Loralee Larios. Loralee’s article, Where and how to restore in a changing world: a demographic-based assessment of resilience, is published in the Journal of Applied Ecology.

The unprecedented rates at which natural systems have been altered have prompted a parallel increase in restoration efforts to recover biodiversity and ecosystem functioning. These restoration efforts are regularly nested within a larger mandate to conduct actions that enhance the resilience of a system to future environmental threats. Even though resilience has become an essential goal in environmental policy and management, its application remains largely heuristic as we continue to struggle with how best to assess and measure it.

Ecological resilience refers to the likelihood that a system will shift to another type of ecological community after a perturbation (i.e., regime shift). Some ecological processes (i.e., feedbacks) can maintain a system in a given state, but managing for a resilient landscape requires knowledge of these feedbacks and is further complicated because feedbacks can change between states resulting in hysteresis where the path to degradation may be very different from that of recovery. Here, we used an integrative approach to identify the various components that contribute to resilience. We leveraged experimental and observational studies to parameterize population models to forecast community dynamics that identified multiple stable communities and hysteresis.

We conducted our study within the context of California grasslands – a highly invaded system that is becoming increasingly vulnerable to more invasion due to atmospheric nitrogen (N) deposition.  We pooled data from various studies focused on grasses that characterize the native and invaded grassland areas – Stipa pulchra and grasses of the genus Avena, respectively- to parameterize demographic models. As a subset of these studies were conducted along an experimentally created resource gradient of low, moderate and high N levels we were able to parameterize these models under three different N scenarios.

We observed that both Avena and Stipa grasslands could persist but this persistence was dependent on N inputs. Under low N, Stipa dominated a community regardless of whether it or Avena was the initial resident. At moderate N levels, we observed strong priority effects that resulted in hysteresis dynamics, where Stipa resisted invasion by Avena and Avena maintained dominance even though Stipa could establish at low abundances (~3 adults/m²). At high N, communities converged to dominance by Avena regardless of starting conditions, with an average of just 1 Stipa adult being able to establish.

By integrating empirical data into demographic models we were able to project dynamics that can be used to prioritize management efforts of conservation of native grasslands and restoration of exotic grasslands in the face of increasing N deposition. Our results indicated that in low resource areas, it is best not to intervene and allow the natural recovery of native grasslands as ecological processes that promote the abundance of the native grass are still intact. Additionally, we argue that restoration should be employed selectively in areas of intermediate N that have the potential to maintain the native state. Moreover, our results suggest that management of existing conservation areas should vary depending on whether or not a target area is approaching a transition. Finally, we find that restoration goals should vary given the environmental context, as some efforts can be risky (i.e., invaded areas receiving high N inputs) due to the high resilience of the degraded state.

This combined collaborative approach that capitalizes on both field data and the quantitative synthesis of these data can inform how best to allocate limited management resources as baselines shift, where natural recovery is possible, where best to allocate active restoration efforts, and where native remnants may be most vulnerable. Restoration will increasingly be required to incorporate more quantitative modelling, as these approaches are well-suited to forecast complex dynamics under future environmental change. This approach will require strong collaborative efforts between land managers and scientists. Thus moving forward, as scientists, we must promote partnerships at the local level to collaborate with monitoring program design as well as process-based experimental assessments of mechanisms.

Where and how to restore in a changing world: a demographic-based assessment of resilience, is a part of the Special Feature, Toward prediction in the restoration of biodiversity and available in read in the Journal of Applied Ecology.

Read other blogs from the series:

• Toward prediction in the restoration of biodiversity by Lars Brudvig
• Every restoration is unique by Katharine Stuble 
• A common currency for connecting the goals of restoration: plant traits can help us understand how plant communities form and help ecosystems function by Chad Zirbel
• Variation in soil microbial communities leads to variation in plant communities by Jonathan Bauer
• Tree islands for tropical forest restoration: the outlook is rosy after 10 years by Leighton Reid (shared from Natural history of Ecological Restoration)
• Finding missing branches: Phylogenetic patterns of plant community diversity in restored and remnant tallgrass prairies by Rebecca Barak