Search

Vegetation phenology and productivity play a crucial role in surface energy balance, plant and animal distribution, and animal movement and habitat use and can be measured with remote sensing metrics including start of season (SOS), peak instantaneous rate of green-up date (PIRGd), peak of season (POS), end of season (EOS), and integrated vegetation indices. However, for most metrics, we do not yet understand the agreement of remotely sensed data products with near-surface observations. We also need summaries of changes over time, spatial distribution, variability, and consistency in remote sensing dataset metrics for vegetation timing and quality. We compare metrics from 10 leading remote sensing datasets against a network of PhenoCam near-surface cameras throughout the western United States from 2002 to 2014. Most phenology metrics representing a date (SOS, PIRGd, POS, and EOS), rather than a duration (length of spring, length of growing season), better agreed with near-surface metrics but results varied by dataset, metric, and land cover, with absolute value of mean bias ranging from 0.38 (PIRGd) to 37.92 days (EOS). Datasets had higher agreement with PhenoCam metrics in shrublands, grasslands, and deciduous forests than in evergreen forests. Phenology metrics had higher agreement than productivity metrics, aside from a few datasets in deciduous forests. Using two datasets covering the period 1982–2016 that best agreed with PhenoCam metrics, we analyzed changes over time to growing seasons. Both datasets exhibited substantial spatial heterogeneity in the direction of phenology trends. Variability of metrics increased over time in some areas, particularly in the Southwest. Approximately 60% of pixels had consistent trend direction between datasets for SOS, POS, and EOS, with the direction varying by location. In all ecoregions except Mediterranean California, EOS has become later. This study comprehensively compares remote sensing datasets across multiple growing season metrics and discusses considerations for applied users to inform their data choices.

Establishment and spread of non-native species following wildfires can pose threats to long-term native plant recovery. Factors such as disturbance severity, resource availability, and propagule pressure may influence where non-native species establish in burned areas. In addition, pre- and post-fire management activities may influence the likelihood of non-native species establishment. In the present study we examine the establishment of non-native species after wildfires in relation to native species richness, fire severity, dominant native plant cover, resource availability, and pre- and post-fire management actions (fuel treatments and post-fire rehabilitation treatments). We used an information-theoretic approach to compare alternative hypotheses. We analysed post-fire effects at multiple scales at three wildfires in Colorado and New Mexico. For large and small spatial scales at all fires, fire severity was the most consistent predictor of non-native species cover. Non-native species cover was also correlated with high native species richness, low native dominant species cover, and high seeded grass cover. There was a positive, but non-significant, association of non-native species with fuel-treated areas at one wildfire. While there may be some potential for fuels treatments to promote non-native species establishment, wildfire and post-fire seeding treatments seem to have a larger impact on non-native species. [ABSTRACT FROM AUTHOR]