Road-stream crossings represent a significant source of habitat fragmentation for
global aquatic ecosystems, yet integrative conservation planning frameworks are lacking
for most regions. We describe a connectivity conservation planning framework that
draws on recent advances in the fields of surveying, modelling, and optimizing removal
of crossings that block aquatic organism passage. We demonstrate this framework with
a case study involving 1200 crossings surveyed in Florida, USA in which barrier severity
was quantified on a continuous scale from 0 (complete barrier) to 1 (no barrier).
Using field surveys, we built a boosted regression tree (BRT) model that linked barrier
severity to 44 landscape variables representing natural (e.g., stream size) and anthropogenic
(e.g., land use) stream conditions. We used a recently developed optimization routine
to conduct two scenarios, including (1) surveyed crossings only and (2) surveyed crossings
plus crossings modelled at 5545 unsurveyed locations. The BRT model explained 54%
of variation in barrier severity scores and showed that the most severe barriers occurred
on small, high-elevation streams draining urban and agricultural catchments. Estimates
of connectivity gains following remediation were 5.3-times lower when unsurveyed locations
were included in the optimization, suggesting inclusion of unsurveyed sites is critical
for conservation planning. Results from this framework can be used over short (e.g.,
planning immediate barrier mitigation) and long (e.g., planning future field surveys)
time horizons to benefit aquatic connectivity conservation. The survey protocol and
modelling methods used here, combined with global datasets on stream conditions, can
be applied to benefit connectivity planning in other regions.