Plot scale

For the small-scale experiments, four main monitoring areas were set up within each of the two sub-catchments at each of the three field sites. These areas were selected to have similar vegetation composition (heather dominated) and slope ranges (5-20°) but different aspects (about 45 degree differences) (see Figure in Expt design section) and are treated as ‘blocks’ within the statistical analysis (with ANOVA) when data for all three field sites are considered together for treatment effects. Within each block, five transects are surveyed manually so that the location of each (5 x 5 m) monitoring plot within a block can be matched for peat depth. Three of the transects contain additional monitoring stations to assess slope and aspect impacts on water table, runoff DOC, peat depth, bulk density and temperature, suitable for model parameterisation for later up-scaling (i.e. topographic effect).

   Mossdale catchment after burning (some burn areas are visibly black)   Mossdale catchment after mowing (some mow areas are visibe)    Whitendale "do nothing" plot within a mown area

We would have thought that the heavy machinery would have caused compaction of the surface peat layers - we actually noticed an about 20 cm sinking of our plastic tube dipwells. However, the peat bounced back up (and so did the rhizon DOC water samplers). We therefore checked the peat depth manually after mowing (2013) and one year later (2014) determined any bulk density changes by repeated coring at mown, burnt and uncut plots to compare bulk densities at the plot level to pre-treatment values. There were no differences in bulk densities for only the mown treatments over the top 50 cm at either sites (the below graph shows data for the three sites: D1 0-5 cm; D2 10-15 cm; D3 20-25 cm; D4 40-45 cm). In fact, Nidderdale and Mossdale showed a general increase in bulk densities for burnt and uncut plots at the surface as well as for mown plots. These differences might represent sampling artefacts (i.e. defining the peat surface is difficult) or different moisture regimes at the sampling times and thus shrinkage/expansion issues. However, Whitendale was virtually unchanged.

Comparison of peat depth across all three sites as control (C) burnt vs. mown treatment (T) and per treatment (T1-T5) and per block (1-4); DN is the uncut comparison. This also did not show any detectable peat compaction after mowing.

Although we could not detect any peat compaction due to mowing with heavy machinery (BD or depth) we did detect impacts of mowing on the plot level micro-topography. We measured about 50 grid points across each 5 x 5 m plot for their offset against the mean peat surface (outside the plot) and compared the STDEV between treatments. The STDEV comparison showed that burnt (C) and uncut (DN) plots were similar whereas mown plots showed a reduced mean STDEV. In addition, a blox plot (shown for Whitendale) revealed the same pattern, note particularly the larger offset ranges (quartiles) for C and DN versus all mown treatments (M: with or without brash removed (BR) or Sphagnum (Sp) added), and M treatments all showing a lower median offset (reduced hummocks). The similarity between C and DN treatments support this reflects a real mowing effect and not differences in plot location. Note: a zero offset was applied as offset values range from positive to negative (below mean peat level).