• Greater restored moss cover decreased peat burn severity. • Deep vs shallow harvesting depth drove divergent post-fire soil water conditions. • Shallow harvest increased suitable conditions for Sphagnum establishment. • Deep harvest lowers the risk of subsequent peat ignition. • Deep harvest likely to promote longer-term carbon sequestration due to fewer fires. Peatland disturbances can disrupt the ecohydrological functions that sustain net carbon sequestration in peatlands. Anthropogenic disturbances, such as peatland drainage and harvesting, are often followed by peatland restoration that aims to return the carbon sink function. This is typically achieved by raising the water table and re-establishing keystone Sphagnum moss species. However, with an increasingly uncertain climate and intensifying land-use changes, the potential for multiple disturbances (such as co-occurring wildfires, drainage, and harvesting) to disrupt the ecohydrological feedbacks that support peatland function is increasing. Yet, few studies investigate the ecohydrological trade-offs induced by multiple disturbances in peatlands. To elucidate the complexities of multiple disturbances and restoration on Sphagnum re-establishment and wildfire potential, we studied a Deep and Shallow harvested area in a drained and restored peatland in southern Ontario, Canada that experienced a wildfire in 2012. Harvesting depth did not significantly increase the bulk density of the upper 32 cm of exposed peat, but the shallower harvest depth did significantly increase the depth of burn (DOB) due to the more varied remnant topography. The difference in topography of the shallower harvested area increased peat carbon losses (16.5 kg C m −2 ) from the wildfire relative to the deeper harvest area (15.1 kg C m −2 ). The difference in post-fire peat hydrophysical properties of the Deep and Shallow harvest area drove divergent soil water conditions. In the post-burn peat, the establishment of suitable conditions for the regeneration of Sphagnum mosses was more prevalent at the Shallow harvest areas but the higher soil water retention capabilities of the Deep harvest peat lowered the risk of subsequent peat ignition. This study highlights the complex interactions multiple disturbances have on peatland ecohydrology and that we urgently need to understand these interactions to better manage our shared peatland resources in an increasingly uncertain future.

• The CRHM-created Boreal Hydrology Model performed quite well on simultaneously simulating runoff, snow water equivalent, soil liquid water content and evapotranspiration (ET) with minor parameter calibration. • The basin hydrological variables showed quite different sensitivities to perturbations of precipitation (P) and temperature (T). Annual runoff was more sensitive to rising P than warming T, but annual ET was more sensitive to warming T. • Perturbed P and T had distinctively different influences on the streamflow regime. Increased P enhanced the intra- and inter-annual variabilities of basin runoff, whilst rising T resulted in the inverse changes. • Effects of warming on annual runoff and snow processes could be compensated for to varying degrees by the effects of increases in P. Hydrological processes over and through frozen and unfrozen ground were simulated in the well instrumented boreal forest basin of White Gull Creek, Saskatchewan, Canada using a model created using the flexible Cold Regions Hydrological Modelling (CRHM) platform. The CRHM-created Boreal Hydrology Model was structured and initially parameterized using decades of process hydrology research in the southern boreal forest with minor parameter calibration, and generally produced quite good performance on simultaneously reproducing the measurements of runoff, snow water equivalent (SWE), soil liquid water content and eddy correlation flux tower observations of evapotranspiration (ET) over two decades. To examine the sensitivity of basin hydrology to perturbed climate inputs, air temperature (T) inputs were set up by linear increments in the reference observation of up to +6 ℃, and precipitation (P) inputs were generated by multiplying the reference observed P from 70% to 130%. The model results showed that the basin hydrological variables showed quite different sensitivities to perturbations of P and T. The volume of annual runoff and the annual runoff coefficient increased more rapidly with rising P, at rates of 31% and 16% per 10% increase in P, but decreased by only 3.8% and 4.7% per 1 ℃ of warming. Annual ET increased rapidly with temperature, by 7% per 1 ℃ of warming and therefore drove the streamflow volumetric changes with warming, but increased only 1% per 10% increase in P. Perturbations of P and T had distinctively different influences on the streamflow regime. Increased P enhanced the intra- and inter-annual variabilities of basin runoff, reduced the relative contribution of winter runoff to annual runoff and increased the relative contribution of summer runoff; whilst rising T resulted in the inverse changes in the streamflow regime. Effects of warming on some hydrological processes could be compensated for to varying degrees by the effects of increases in P. Reductions in the annual runoff volume and runoff coefficient caused by warming up to 6 ℃ could be compensated for by increases of <20% in P. However, the maximum increase in P (+30%) examined could only compensate for the changes in snow processes caused by warming of less than 4 ℃ and snow-cover duration decreases with 1 ℃ warming could not be compensated for by any precipitation increase considered. These results inform the vulnerability of boreal forest hydrology to the first-order changes in P and T and provide guidance for further climate impact assessments for hydrology in the southern boreal forest in Canada.