Monitoring

View article.

We developed a framework for identifying which biotic traits would provide the best initial indication of longer-term target restoration goals and applied the framework to restoration drill-seedings of deep-rooted perennial bunchgrasses (DRPBGs) used to rehabilitate and restore semiarid rangelands threatened by exotic annual grasses (EAGs, e.g. cheatgrass) and the recurrent wildfire that EAGs cause. Initial traits measured included cover, basal diameter, height, and density (#plants/area) of DRPBGs and cover of EAGs and Sandberg bluegrass (Poa secunda, POSE, a disturbance-adapted perennial). The longer-term target objective was ≥25 % DRPBG cover and ≤13 % EAG cover by the 5th year following drill-seedings. Measurements were made on 112 plots spanning 113,000 ha in sagebrush steppe on the Soda wildfire scar, in the Northern Great Basin, USA. Traits of DRPBGs tended to be uncorrelated with one another, thus each was informative in describing vegetation condition. Where DRPBG cover was initially >17 %, it tended to become >25 % by the 5th-year post-seeding. In plots that overcame an initial risk of not meeting the target objective (i.e. <17 % initial DRPBG cover), DRPBG tended be large DRPBGs (>22.8 cm height) and plots also had >7 % cover of POSE. Additional “sets” of initial vegetation traits were also predictive of longer-term restoration success or failure. Restoration drill-seeding of DRPBGs is a key but varied-outcome tool for breaking the exotic grass-fire cycle, and, contrary to a conventional tendency to rely on a limited number of mean traits such as % cover, a suite of biotic traits appears necessary to monitor to reliably know if trials are likely to yield success.

View article.

Large-scale disturbances, such as megafires, motivate restoration at equally large extents. Measuring the survival and growth of individual plants plays a key role in current efforts to monitor restoration success. However, the scale of modern restoration (e.g., >10,000 ha) challenges measurements of demographic rates with field data. In this study, we demonstrate how unoccupied aerial system (UAS) flights can provide an efficient solution to the tradeoff of precision and spatial extent in detecting demographic rates from the air. We flew two, sequential UAS flights at two sagebrush (Artemisia tridentata) common gardens to measure the survival and growth of individual plants. The accuracy of Bayesian-optimized segmentation of individual shrub canopies was high (73–95%, depending on the year and site), and remotely sensed survival estimates were within 10% of ground-truthed survival censuses. Stand age structure affected remotely sensed estimates of growth; growth was overestimated relative to field-based estimates by 57% in the first garden with older stands, but agreement was high in the second garden with younger stands. Further, younger stands (similar to those just after disturbance) with shorter, smaller plants were sometimes confused with other shrub species and bunchgrasses, demonstrating a need for integrating spectral classification approaches that are increasingly available on affordable UAS platforms. The older stand had several merged canopies, which led to an underestimation of abundance but did not bias remotely sensed survival estimates. Advances in segmentation and UAS structure from motion photogrammetry will enable demographic rate measurements at management-relevant extents.

View article.

Adequate numbers of replicated, dispersed, and random samples are the basis for reliable sampling inference on resources of concern, particularly vegetation cover across large and heterogenous areas such as rangelands. Tools are needed to predict and assess data precision, specifically the sampling effort required to attain acceptable levels of precision, before and after sampling. We describe and evaluate a flexible and scalable process for assessing the sampling effort requirement for a common monitoring context (responses of rangeland vegetation cover to post-fire restoration treatments), using a custom R script called “SampleRange.” In SampleRange, vegetation cover is estimated from available digital-gridded or field data (e.g., using the satellite-derived cover from the Rangeland Assessment Platform). Next, the sampling effort required to estimate cover with 20% relative standard error (RSE) or to saturate sampling effort is determined using simulations across the environmental gradients in areas of interest to estimate the number of needed plots (“SampleRange quota”). Finally, the SampleRange quota are randomly identified for actual sampling. A 2022 full-cycle trial of SampleRange using the best available digital and prior field data for areas treated after a 2017 wildfire in sagebrush-steppe rangelands revealed that differences in the predicted compared with realized RSEs are inevitable. Future efforts to account for uncertainty in remotely sensed−based vegetative products will enhance tool utility.