appendix c. the impact of ditch blocking on blanket

of 37

Appendix C. The impact of ditch blocking on blanket peatland hydrology Joseph Holden, Sophie M. Green, Andy J. Baird, Richard P. Grayson, Pippa J. Chapman, Chris Evans, Michael Peacock To be cited as: Holden, J., Green, S.M., Baird, A.J., Grayson, R.P., Chapman, P.J., Evans, C., and Peacock, M. 2016. The impact of ditch blocking on blanket peatland hydrology. Appendix C in: Green, S.M., Baird, A.J., Evans, C., Ostle, N., Holden, J., Chapman, P.J., and McNamara, N. Investigation of Peatland Restoration (Grip Blocking) Techniques to Achieve Best Outcomes for Methane and Greenhouse Gas Emissions/Balance: Field Trials and Process Experiments Final Report. Defra Project SP1202, University of Leeds, Leeds.

1 Introduction Carbon (C) cycling processes in peat soils depend on the hydrological status and regime of the peatland (Holden, 2005; Limpens et al., 2008). Globally, peatlands store over a third of all soil carbon (Yu et al., 2010). In the UK, blanket peat accounts for 87 % of peatland cover (Baird et al., 2009) and exists primarily in the uplands, covering mostly gently rolling terrain. Blanket peat depths are typically 1-3 m but can be in excess of 6 m in places. The hydrological regime of blanket peatland tends to be dominated by the movement of water at or close to the surface due to saturation of the peat and a low hydraulic conductivity in all but the shallowest layers of peat (Evans et al., 1999; Holden and Burt, 2003a; Holden and Burt, 2003b; Price, 1992). Thus, it has often been assumed that the surface topography alone can be used to determine the water catchment area for blanket peatland streams. However, concentrated subsurface flow along channels (pipes) has been shown to account for 10-14 % of stream discharge in blanket peatlands (Holden and Burt, 2002; Smart et al., 2013). These pipes may undulate through the peat profile (Holden, 2004) and form branching networks, and their course may transport water from source areas outside of the surface topographic boundary (Jones, 1997). Thus the catchment area for small blanket peatland systems may not necessarily coincide with their surface topography, but such an idea has never been tested with flow measurements before. Many peatlands in the UK were artificially drained between the 1950s and 1980s to support agricultural demand in areas of marginal productivity (Baldock et al., 1984; Green, 1974), for commercial forestry (Holden et al., 2007), to aid peat extraction for horticulture and energy production, and because of the perception that peatland drainage could alleviate flood risk (Acreman and Holden, 2013; Holden et al., 2006; Newson, 1992). Drainage of blanket peat in the UK has been relatively widespread and has most commonly taken the form of open ditches dug to around 50 cm depth and 50 cm width and spaced at distances of between 5 m and 50 m. In response to concerns about biodiversity, erosion, carbon storage and potentially exacerbated flood risk, many UK peatland ditch networks are being blocked as part of peatland restoration measures (Armstrong et al., 2009; Parry et al., 2014). Many techniques are used for ditch blocking; particularly common is the installing of peat dams at intervals of several metres along the course of the ditch (Armstrong et al., 2009). Sometimes ditches are partially infilled using peat from the areas between the ditch channels, the result being a much shallower channel on which a vegetation cover develops (Parry et al., 2014); this method is called reprofiling. Ditch blocking and reprofiling will influence the hydrological flowpaths on and in blanket peatlands. Depending on the success of the dams constructed during peatland ditch blocking, water can pond behind dams to create pools (Beadle et al., 2015) and then excess water can be forced out of the ditch channel across the surrounding peat surface to follow topographic drainage routes (Holden, 2006). These routes may differ from the routes that existed before blocking (Lane and Milledge, 2012). Open blanket peatland ditches tend to capture saturation-excess overland flow and subsurface flow from upslope and, if the ditch is orientated in a direction even slightly offset from the natural downslope direction, the peat on the downslope side of the ditch can be expected to receive less flowing water than it would in the absence of the ditch. The reduction in water flowing past a given point on the landscape due to upslope diversion along a ditch channel can be expected to cause a lowering of water tables (Holden et al., 2011), reductions in overland flow (Holden et al., 2006) and possibly peat subsidence or compression (Price and Schlotzhauer, 1999).

of 37

Even when ditches are orientated in a downslope direction there could be local effects of the ditches on peatland water tables and therefore on the propensity to saturation and the amount of saturation-excess overland flow. This is because the ditch will tend to have a lower water level within it than the relative topographic height of the water-table within the surrounding peat and hence there will be a hydraulic gradient towards the ditch. Indeed, this may be another reason why surface topography alone may not be suitable for determining the catchment area of peatland streams or ditches in small catchments because the slope of the hydraulic gradient may not necessarily match the topographic slope. Such differences between the hydraulic gradient and the surface slope orientation may be more likely to have an effect on the catchment water budget calculation at the scale of small catchments around ditches. Further work is required to measure hydraulic conductivity in blanket peatlands. Results to date have shown that while hydraulic conductivity can be high very near the peat surface and above pipes (e.g., Cunliffe et al., 2013), for most of the peat profile it is very low (Holden and Burt, 2003a; Lewis et al., 2011). Where the hydraulic conductivity is low, hydraulic gradient effects on water flow into ditches running downslope may be minor. There have been few studies of the hydrological impact of ditch blocking in situations where peatland ditches run predominately downslope. Most studies have examined water movement and water-table behaviour in and around blocked ditches where the ditches are orientated across the slope (e.g., Gibson et al., 2009; Holden et al., 2011; Wilson et al., 2010; Worrall et al., 2007) or where the site is virtually flat (e.g., Haapalehto et al., 2011; McCarter and Price, 2013). However, there are large areas of blanket peatland for each of these conditions (downslope, cross-slope, almost flat) in the UK. The hydrological data reported here serve to support calculations and understanding of the gaseous and waterborne carbon fluxes from the study site (see Appendices D and E). Data on ditch flow, overland flow and water tables are analysed and presented. Flows and water-table behaviour around dammed, dammed and reprofiled (henceforth called reprofiled) and open ditches will be compared.

2 Methods The overall monitoring set up at the study site, vegetation, peat and climate conditions are described elsewhere in this report (see section 2 of the main part of the report). Here we focus on the hydrological measurements.

2.1 Measurements

2.1.1 Discharge Discharge from the 12 ditches was measured using 22.5o v-notch weirs and WT-HR 1000 water-height loggers (TruTrack Ltd, Christchurch, New Zealand) (installed 24th June 2010). The water-height loggers were set up to record at 15-minute intervals, with these logged values being an average of 60 s point samples. Each weir was manually calibrated, in two stages. First, the TruTrack was calibrated, so the logged water heights were known to be accurate. Secondly, for a known TruTrack-recorded water level, the v-notch was calibrated by collecting manual discharge data. Discharge of overland flow on the areas between ditches was sampled. The overland flow from both sides of each study ditch (see Figure 2.1 below) was channelled into one central overland flow weir box per ditch using ultraviolet-stable polyvinyl chloride (PVCu) soffit boards pushed into the peat to a depth of 3 cm. The soffit boards varied in length running from the approximate mid-point between ditches towards the ditch. Overland flow channelled from one side of the ditch crossed over the ditch via a PVC pipe before entering the overland flow weir box near the ditch. The overland flow weir boxes (installed July 2011) were also fitted with TruTrack WT-HR 500 loggers. The weir boxes were calibrated in the same way as the ditch v-notch weirs. Discharge was logged at 15-minute intervals (averaged from 60 s point intervals). Discharge

of 37

from the overland flow weir boxes occurred downslope of the main ditch weir. This was to ensure that the capture and redirection of overland flow did not affect the flow into the gauged sections of the ditches. Because the soffit boards were inserted to a depth of 3 cm, overland flow in our analysis refers to flow at the surface and within the upper 3 cm of the peat profile. There were occasional logger failures over the course of the study. Unreasonable values from the ditch flow and overland flow records caused by, for example, icing up, or occasional one-off erroneous readings from the pressure sensor were removed before data analysis. Where there was a data gap of two points or less

(2 15-min intervals) in the automated record the values were infilled using linear interpolation. Otherwise the data gaps were retained and reported as missing values to be taken into account when interpreting the total water flux.

2.1.2 Water-table depths Eighty-three dipwells for monitoring water-table depth across the study site were installed. Twenty-four of these were fitted with automatic water-level recorders (see below), while water levels in the rest of the dipwells were recorded manually during site visits (approximately once every three weeks in summer and every six weeks in winter). The automated dipwells comprised high-density polyethylene (HDPE) pipes, with an outside diameter (o.d.) of 32 mm, a length of 1000 mm and a 3.5 mm wall thickness. These pipes were perforated with numerous 2 cm horizontal, 0.3 mm wide slits spaced at intervals of 5 mm. The automated wells were located midway between each ditch and also within each ditch to measure water level within the ditch (or water-table depth below the ditch floor). They were fitted with either a WT-HR 1000 water-height data loggers (TruTrack Ltd), or Divers (DI240, 5 m, Schlumberger Water Services, Delft, The Netherlands), which measured and recorded water tables at 2-hr intervals. Each data logger was manually calibrated, and thereafter checked and cleaned throughout the project at regular intervals. Manually

gauged dipwells were made from 32 mm o.d. 1000 mm (PVC) pipe with a 3.5 mm wall thickness. The tubes were perforated with 8-mm diameter holes drilled at 100 mm intervals along four lines running lengthwise along the pipe, with holes in each line offset from those in neighbouring lines by 50 mm. The manually recorded dipwells were located at 2 m from each ditch both on the eastern and western sides (coded x.2E or x.2W, where x is the ditch number) adjacent to the automated dipwell within the ditch (coded x.0) and in line with the automated dipwells that were installed half way between ditches (coded x.mid). The installation of the dipwells was completed in August 2010. The wells were removed in early February 2011 prior to the on-site damming and reprofiling work on the ditches. Re-installation of all dipwells was completed by 23rd February 2011. All dipwells were tested for their response time by measuring the recovery of well water levels in response to a sudden withdrawal of water. The time for 90% recovery ranged from a few seconds to 120 minutes. Thirty-five manually-recorded dipwells were installed in the first week of June 2011 associated with a set of flux-chamber collars (see Green et al. (2016); Appendix D). These were located within the ditch channel and 1 m and 3 m west of each ditch (coded CSx 0 m, CSx 1 m or CSx 3 m where x is the ditch number). In Ditch 5 the flux-chamber collar was coincident with the automatic dipwell; hence, no additional manual dipwell was installed at that point. The flux-chamber collar dipwells were specifically installed so that CH4 and CO2 fluxes derived from the flux chamber measurements could be directly related to local water-table depth as described in Appendix D (Green et al., 2016). These hydrological data are, nevertheless, analysed and discussed below.

2.2 Calculations Total discharges of water were calculated for each weir in m3. These data were also converted into mm of discharge (which can be related to total precipitation) by determining the surface topographic catchment area of each weir. Figure 2.1 below shows the surface-derived catchment area for each weir calculated from LiDAR digital elevation model (DEM) data for the study site provided by the National Trust. However, the damming of ditches may lead to changes in catchment areas for adjacent ditches if water is shed from

of 37

dammed ditches into nearby ditches. While this effect is minimised in the study system because many of the ditches follow an approximately downslope direction, there may still be some spillage effects from dams in ditches where the ditches do not run directly downslope. To examine the maximum theoretical effect of such changes in surface hydrological flowpaths, a DEM was created where it was assumed that all of the blocked ditches had been completely infilled with peat and thus no longer existed as topographic features in the landscape. The four open or control ditches were, however, left in place in the DEM. The catchment areas for ditch weirs under this infilling scenario are shown in Figure 2.2. There would also be some modifications to the surface catchment areas for overland flow collectors under such a scenario. The catchment areas for both scenarios (scenario 1: pre-blocking catchment areas; scenario 2: catchment areas calculated for an assumed situation involving the complete infilling of dammed and reprofiled ditches) are shown in Table 2.1 and both are used when calculating mm of discharge from each ditch or overland flow weir box. Note that the infilling scenario is an extreme one and unlikely to be fully met within our study system during the initial few years after management intervention. However, by calculating the area-weighted discharge from the weirs and its change over time it will be possible to evaluate whether any such redistribution effects have been occurring. Note that the maximum infilling scenario results in very small catchment areas for the weirs in Ditches 3, 4, 5, 11 and 12 (Table 2.1). In the case of Ditches 3, 4 and 5, however, there is a large theoretical rise in the catchment area for the overland flow weirs. For Ditches 10, 11 and 12 the overall catchment area for both overland flow and ditch flow weirs are substantially reduced in the infilling scenario largely due to downslope flow towards the west away from the weirs. The overall area of the monitored hillslope that captures both overland flow and subsurface flow (i.e., the ditch weirs) is 24 % smaller in the infilled scenario. Hence, if ditch blocking at the site is effective we would expect to capture less total flow in the ditch weirs and overland flow weirs because about 24 % of the water should be diverted away from blocked ditches to follow the topographic gradient of the site in a direction that partly by-passes the weirs. The catchment area supplying overland flow weir boxes in the ditch infilling scenario is about three times greater than the one with open ditches (Table 2.1).

of 37

Figure 2.1 Surface-derived catchment areas for each ditch and each overland flow (OLF) weir box. The surface catchment area for each ditch weir is shown by the large shaded areas, while the overland flow weir box catchment areas at the lower (northern) end of each ditch system are shown by the green and pink shading as highlighted in the legend.

of 37

Figure 2.2. Theoretical catchment areas (shaded areas) for each ditch and overland flow weir box assuming that dammed (4, 5, 10, 12) and reprofiled (1, 3, 8, 11) ditches are completely infilled, while open ditches (2, 6, 7, 9) remain. The original catchment areas for the ditch weirs (from Figure 2.1) are shown with a red boundary for comparison Figure 2.2 highlights that Ditches 6 and 7, which were open ditches, are good controls because their theoretical catchment areas are hardly affected under the maximum infilling scenario. In this report, calculations to compare blocked ditch flow and associated overland flow and water-tables to the open control treatments are typically made with reference to catchment 6. If the particular dataset being analysed from catchment 6 had any major problems (such as missing data due to data logger malfunction) then comparisons were made to the matching dataset from catchment 7. Time series data are presented to check for trends over time because there may be hydrological responses to ditch blocking that evolve relatively slowly rather than in the immediate aftermath of ditch blocking. While the focus of the analysis (and the bulk of the data collected) in the project relates to testing blocking effects on carbon fluxes compared to unblocked controls, where possible we take advantage of the data collected before blocking occurred in order to strengthen our control-treatment analysis and

of 37

interpretations. When data have been evaluated on an annual basis, the data are treated in full years from 1st March (2011/12, 2012/13, 2013/14, 2014/15) because ditch blocking took place in February 2011. Data from before March 2011 are included, where relevant and available, and tend to run for the 6 months from August 2010 to January 2011 inclusive, covering both summer and winter periods. Table 2.1. Surface topography derived catchment areas (m2) of the study weirs for two scenarios.

Ditch Ditch flow Overland flow

Assuming all open ditches

Assuming treatment ditches act as if they are infilled

Assuming all open ditches

Assuming treatment ditches act as if they are infilled

1 2942 2499 100 69

2 1950 2537 229 322

3 2426 105 291 777

4 1350 19 61 527

5 969 38 43 1639

6 1462 1494 79 117

7 1227 1340 45 54

8 1195 823 280 50

9 1642 1997 55 25

10 2142 1329 161 103

11 1541 4 63 713

12 1311 40 108 37

Total 20157 12225 1515 4433

For some of the dipwell and ditch flow data, calculations were performed to determine the relative impacts of ditch blocking compared to the control ditches using both before and after blocking datasets. Using comparisons between controls and treatments for both before and after blocking periods helps deal with problems around differences in water-table depth between treatments resulting solely from weather conditions experienced in a given year (wetter years and drier years). For example, for each particular dipwell location (e.g., 2 m west of the ditch) the mean difference between water-table depth was determined relative to the control at open ditch 6 ((control water-table depth) minus (study ditch water-table depth)). This mean ‘offset’ was calculated for the period before any management interventions took place on site. The calculation was repeated for study year periods 1 March to end of February in subsequent years. The former offset value was subtracted from the latter offset values. If the resulting number for a given year was greater than zero this suggested a relatively ‘positive’ change in the treatment water-table depth compared to the control (i.e., the water tables had become shallower (closer to the surface) relative to the control).

3. Results

3.1 Ditch flow The time series hydrographs are shown in Figure 3.1. For many of the ditches that were dammed or reprofiled in February 2011 there was an immediate effect on the discharge regime. The one exception was Ditch 3 where the effect of blocking on the discharge pattern is not so obvious. However, 30 % of the data logger record for Ditch 3 was lost during the August 2010 to January 2011 (inclusive) pre-blocking phase. In consequence many of the higher peaks in flow for the earlier part of the record that were observed in the other ditches occurred during periods when the data are missing for the Ditch 3 record. The ditches that were left open (2, 6, 7, and 9) show larger peak flows throughout the study period through to February 2015 than for the post-blocking phase of the other ditches.

of 37

Figure 3.1. Discharge record for the study ditches, 1st August 2010 to 25th February 2015. Management interventions on the ditches took place in February 2011 and hence there is a gap in all ditch flow records for that month – dashed lines indicate the timing of the interventions for affected ditches (continued on next page).

of 37

Figure 3.1. (continued) Annual totals for ditch flow are presented in Table 3.1 through to end of February 2015. There are very large differences in total flows between ditches, with the largest total flows after February 2011 occurring in the open ditches (2, 6, 7, and 9). The total flow values from Ditch 2 appear to be very large across the

of 37

study but particularly in 2012/13. There may, therefore, be some problems with the discharge rating curve for the Ditch 2 weir at the high end (where manual calibrations are not common) because 2012/13 had many high-flow events. However, the logger at this site suffered regular failures and was out of action for a large proportion of the time. These data are therefore less reliable and Table 3.1 provides overall totals in the bottom rows that both include and exclude Ditch 2 and its catchment area. The penultimate row of Table 3.1 includes some adjusted estimates of total annual flow from the study area taking account of the proportion of missing data points in the record across all weirs. Such adjustment

involves multiplying the existing value by (1 + ) where is the proportion of missing data (e.g., if 5 % are missing then the total would be multiplied by 1.05). The bottom row of the table provides estimates of runoff efficiency for the study area as a whole (area-weighted discharge expressed as a percentage of precipitation) reported using both catchment area estimates (the before and after blocking estimates – scenarios 1 and 2) for each ditch. For the period before ditch blocking, the runoff efficiency from the site as a whole is estimated to be 82%. However, during this first period of the study two ditches (4 and 9) appear to produce more aerially-weighted discharge than rainfall. After the management intervention period, when using area-weighted discharge based on the original surface-derived catchment areas for each ditch (scenario 1), it appears that the discharge is greater than precipitation in some of the ditches for some of the years, but more so in the last two years of the study. For the overall hillslope, however, the total area-weighted discharge (in mm) is less than rainfall in all years assuming the original catchment area applies. Notably there was an increase from the 2011/12 low value of 32 % in the overall runoff efficiency so that in the final two years of the study the efficiency was 89 % and 71 % respectively, which is close to that observed in the period before ditch blocking. When using the modified (and therefore reduced) overall catchment area (assuming infilled ditches), the runoff efficiency for the whole hillslope in 2013/14 and 2014/15 was calculated at > 100 %.

of 37

Table 3.1. Total annual water fluxes from the ditch weirs (continued on next page).

m3 of water

mm of runoff assuming original catchment area mm of runoff assuming infilled ditch catchment area

Year Before 2011/12 2012/13 2013/14 2014/15 Before 2011/12 2012/13 2013/14 2014/15 Before 2011/12 2012/13 2013/14 2014/15

Rainfall 1238 2255 2409 1786 1888 1238 2255 2409 1786 696

Ditch 1 (reprofiled) 1747 391 534 659 847 594 133 181 224 288 717 160 219 270 348

Ditch 2^ (control – open) 2251 2167 14066 3448 2690

1154 1111 7213 1768 1379 930 895 5812 1425 1111


Ditch 4 (dammed) 4328 2571 4266 8662 1985 3206 1904 3160 6416 1470 10183 6049 10038 20381 4670

Ditch 5 (dammed) 1090 458 1855 2606 4117 1125 473 1914 2689 4249 677 285 1151 1617 2556

Ditch 6 (control – open) 1631 3864 7053 4692 3978 1115 2643 4824 3209 2721 1091 2587 4721 3141 2663

Ditch 7 (control – open) 1065 2554 3618 4413 5570 868 2081 2949 3596 4539 796 1907 2702 3295 4160


Ditch 9 (control –open) 3164 1087 2102 2575 2583 1927 662 1280 1568 1573 1457 501 968 1186 1189

Ditch 10 (dammed) 1218 558 1171 1013 1529 568 260 546 473 714 980 449 942 815 1230


Ditch 12 (dammed) 182 2 5 10 28 139 1 4 7 21 1431 15 40 77 219

Total 19731 14857 36426 31222 26220 979 737 1807 1549 1301 1346 1014 2486 2130 1789

Total (not including Ditch 2) 17480 12690 22360 27773 23530 960 697 1228 1525 1292 1429 1037 1828 2270 1923

Adjustment for missing data 18518 13076 24507 28873 23853 1017 718 1340 1586 1344 1513 1068 1994 2361 2000

Overall runoff coefficient, % 82 32 56 89 71 122 47 83 132 106

*Data for 1st Aug 2010 to 31st Jan 2011 One shade lighter indicates 5-10% data series missing Lightest shading indicates 10-30% data series missing White cell indicates > 30% data series missing ^Ditch 2 flow data are unreliable – see text.

of 37

Table 3.1. (continued – the part of the table below appends to the right of the part shown above) The proportion of time when no flow was occurring in the ditches varied markedly between ditches, with Ditches 1, 9, 11 and 12 having the longest dry periods. The proportion of zero flow time declined for each full year of the study from 2011/12 to 2013/14 in all ditches except the open ones and Ditch 10 (Table 3.2). Further declines (i.e. to produce the smallest values since the ditch blocking took place) in dry flow duration were found for Ditches 1, 8, 10 and 12 for the final year of the study, although dry flow periods were shorter in 2014/15 for control ditches 7 and 9 compared to the other years since ditch blocking took place. Overall, flows were more continuous from ditches in later years of the study. There appeared to be a large reduction in dry periods at the Ditch 10 weir after ditch blocking compared to the period before blocking suggesting that pooled water (water in pools behind dams) upslope was able to slowly seep out of this drain system for long periods after rainfall. This hydrological behaviour is confirmed by less steep flow duration curves for Ditch 10 in the years after ditch blocking (Figure 3.2). Table 3.2. Proportion of time (%) when flows < 0.1 mL s-1 occurred at the ditch weir.

Ditch Before 2011/12 2012/13 2013/14 2014/15

1 (reprofiled) 50.9 80.0 63.2 33.9 22.2

2 (control – open) 13.1 25.4 1.2 5.2 5.9

3 (reprofiled) 7.2 30.1 17.6 10.1 15.1

4 (dammed) 0.0 8.4 0.5 0.0 0.0

5 (dammed) 15.2 0.2 0.0 0.0 0.0

6 (control – open) 0.0 19.8 1.1 8.7 1.9

7 (control – open) 29.3 24.5 9.1 6.7 6.4

8 (reprofiled) 0.4 36.5 15.1 7.4 6.0

9 (control –open) 33.4 69.7 44.9 32.9 31.6

10 (dammed) 49.6 11.1 1.4 7.0 0.8

11 (reprofiled) 49.6 52.4 35.0 18.1 24.3

12 (dammed) 1.4 99.6 98.1 97.6 97.1

% missing data in time series

Year Before* 2011/12 2012/13 2013/14 2014/15

Rainfall 0.0 0.0 0.0 0.0 0.0

Ditch 1 (reprofiled) 0.0 4.5 0.6 0.2 0.2

Ditch 2 (control – open) 1.6 49.5 12.4 6.2 18.0


Ditch 4 (dammed) 0.2 0.0 0.6 0.3 0.2

Ditch 5 (dammed) 0.0 0.0 0.6 6.2 0.2




Ditch 9 (control –open) 8.7 0.0 0.6 0.3 0.2

Ditch 10 (dammed) 0.0 0.0 0.6 6.4 12.8


Ditch 12 (dammed) 6.9 0.0 15.3 3.4 0.2

Total 5.6 6.9 9.3 4.1 2.8

Total (not including Ditch 2) 5.9 3.0 9.1 4.0 1.4

of 37

Figure 3.2. Flow duration curves for all ditch weirs by year (continued on next page). Normalised discharge has no units as it is discharge divided by mean discharge.

of 37

Figure 3.2. (continued)

Flow duration curves show the nature of the flow regime at a site and are typically plotted on a log-probability scale on the x-axis (as shown in Figure 3.2) so that the details of the less frequently occurring upper (peak) and lower (low flow) parts of the record can be examined. In order to enable comparisons between catchments of different sizes, flow duration data are also typically plotted as normalised data (and on a log10 scale1 since data are often clustered at the lower end for the flow record). The normalising is done by dividing discharge at each time in the record by the overall mean flow for that gauging station. Steep gradient curves on flow duration plots indicate a 'flashy' flow regime (rapid rise and fall of flow and frequent extremes of high flow and low flow) while more gentle slopes indicate a more evenly-flowing ditch flow regime over the study period. At the high flow end of the flow duration curves in particular, but for most of the dataset, the control Ditches 6 and 7 have very similar curves between each year; the curves only really deviate between years for Ditch 6 during low flow conditions. The open Ditches 2 and 9 also show similar shapes to their curves across the years. The other ditches, however, show large differences in the slopes of the curves between years, with more separation at the high end of the curves between years and in particular between the before (black symbols) and after blocking periods. The very gentle gradient curves for Ditch 5 after blocking indicate a change to more continuous flow all year as also outlined in Table 3.2. Interestingly, this ditch appears to have a less and less flashy regime year on year as indicted by the progressively less steep gradient curves for each year in the record. The weir at Ditch 5 may also have a very large increase in catchment area (and consequent discharge totals) associated with spillage of water from other blocked drains (Table 2.1). However, it appears that this extra flow (Table 3.1) is mainly associated with slower water release and more enhanced baseflows, rather than higher storm peaks even during the wettest rainfall events (Figure 3.2).

1 Here, the normalised flow data are logged and plotted on a normal scale which is the same as plotting the unlogged data on a log

scale.

of 37

Given the large uncertainty in individual ditch catchment areas after the management intervention it was deemed useful to examine changes in the volume of water produced at each weir per unit of rainfall (Figure 3.3). The total flow passing the ditch weirs declined from 14.96 m3 per mm of rainfall to only 5.80 m3 per mm in the first year after blocking compared to the period before blocking. When only ditches that were blocked are considered that figure was 9.39 m3 mm-1 dropping to 2.30 m3 mm-1. Considering only the period after ditch blocking, most of the ditches experienced a significant increase (p < 0.05) over time (using month since blocking as the sequential time unit in a linear regression) in the volume of water produced per mm of rainfall (Table 3.3). At control Ditch 6 there was no significant increase in runoff volume per mm of rainfall over the same period. Open Ditch 9 did show a significant trend of increasing discharge per unit of rainfall but this is in line with expectations that water from adjacent blocked ditches would flow into that ditch which may have experienced an increase in catchment area of ~20 % after ditch blocking operations compared to the situation before February 2011 (Table 2.1).

Aug-14Aug-13Aug-12Aug-11Aug-10

400

200

0

Pre

cip

ita

tio

n (

mm

)


16

12

8

4

0Dis

cha

rge

pe

r ra

infa

ll u

nit

(m

³ m

m⁻¹

)Ditch 2

Ditch 6Ditch 7Ditch 9


16

12

8

4

0Dis

cha

rge

pe

r ra

infa

ll u

nit

(m

³ m

m⁻¹

)

Ditch 4



16

12

8

4

0Dis

cha

rge

pe

r ra

infa

ll u

nit

(m

³ m

m⁻¹

)

Ditch 1


Open

Dammed Reprofiled

Figure 3.3. Monthly discharge (m3) per unit of rainfall (mm) for each ditch weir. Table 3.3. Correlation coefficients and the gradient of change over time since March 2011 in monthly discharge per unit rainfall (m3 mm-1). Coefficient and gradient values only shown where p < 0.05. Significant p values shown in bold.

Ditch Correlation coefficient Gradient of change p value

1 (reprofiled) 0.45 0.0081 0.001

2 (control – open) 0.087

3 (reprofiled) 0.40 0.0087 0.007

4 (dammed) 0.190

5 (dammed) 0.66 0.0690 <0.001

6 (control – open) 0.747

7 (control – open) 0.51 0.0571 0.001

8 (reprofiled) 0.71 0.0153 <0.001

9 (control –open) 0.44 0.0227 0.002

10 (dammed) 0.47 0.0193 0.001

11 (reprofiled) 0.29 0.0143 0.048

12 (dammed) 0.47 0.0002 0.001

of 37

Figure 3.4 plots the difference, for each 15-minute point in the time series, between flow at the Ditch 6 weir and flow at some of the other ditches as examples (Ditch 6 flow minus flow in Ditch x). The effect of blocking is clear with a major shift in the offset occurring for blocked ditches from March 2011. The ditches that were left open typically showed a mix of positive and negative offsets to the control along the time series while blocked ditches show a mainly positive offset.

01/8/201401/8/201301/8/201201/8/201101/8/2010

12

6

0

Hou

rly

prec

ipita

tion

(mm

)

01/8/201401/8/201301/8/201201/8/201101/8/2010

3

0

-3

-6

01/8/201401/8/201301/8/201201/8/201101/8/2010

5

3

1

-1

-3

Dis

char

ge

diff

ere

nce

(L s

⁻¹)

01/8/201401/8/201301/8/201201/8/201101/8/2010

31

-1-3

-5

Precipi tation

Ditch 1 (reprofi led)

Ditch 5 (dammed)

Ditch 7 (open)

Figure 3.4 Difference between flow at Ditch 6 weir and flow in other example ditches over time (Ditch 6 flow minus example ditch flow).

3.2 Overland flow Overland flow regularly occurred on the site (Figure 3.5; Table 3.4) and all weirboxes recorded overland flow showing that it was spatially widespread. Note that overland flow data were not collected before ditch blocking took place on site. The overland flow hydrographs are rather flashy, but many weirs recorded flow for extended periods well after rainfall, suggesting that saturation-excess overland flow was the dominant surface flow mechanism. Most of the blocked ditch systems, or those with potential overspilling from other drain catchments, have more frequent overland flow than the unblocked control systems 6 and 7 which both have similar frequencies of overland flow (Table 3.4). The overland flow time series are somewhat different between Ditches 6 and 7, and seem to show different responses between the years (Figure 3.5). Unfortunately overland flow data are unavailable for Ditch 6 from July 2014 onwards. The main differences for the period up to July 2014 appear to be restricted to the winter months where there are large spikes in flow either at Ditch 6 or Ditch 7 and these may be related to surface ice and snow effects on overland flow routing and the function of the weir boxes. Overland flow occurred less frequently at weir box 12 in all study years compared to control weir box 7. Weir box 12 is the only one which theoretically would see a reduced catchment area due to drain infilling and this weir box did experience much less frequent overland flow than any other site. The duration of overland flow increased each year relative to control weir box 7 for weir boxes 1-5 (Table 3.4).

of 37

0.3

0.2

0.1

0.0

01/3/201401/3/201301/3/2012

2.4

1.8

1.2

0.6

0.0

01/3/201401/3/201301/3/2012

0.4

0.3

0.2

0.1

0.0

0.16

0.12

0.08

0.04

0.00

Ditch 1 (reprofiled)D

isch

arg

e (

L s⁻

¹)Ditch 2 (open)

Ditch 3 (reprofiled) Ditch 4 (dammed)

0.12

0.09

0.06

0.03

0.00

01/3/201401/3/201301/3/2012

1.2

0.9

0.6

0.3

0.0

01/3/201401/3/201301/3/2012

0.24

0.18

0.12

0.06

0.00

0.100

0.075

0.050

0.025

0.000

Ditch 5 (dammed)

Dis

cha

rge

(L

s⁻¹)

Ditch 6 (open)

Ditch 7 (open) Ditch 8 (reprofiled)

Figure 3.5. Overland flow, March 2012 to February 2015 for each weir box (continued on next page).

of 37

0.100

0.075

0.050

0.025

0.000

01/3/201401/3/201301/3/2012

0.8

0.6

0.4

0.2

0.0

01/3/201401/3/201301/3/2012

0.4

0.3

0.2

0.1

0.0

0.20

0.15

0.10

0.05

0.00

Ditch 9 (open)D

isch

arg

e (

L s⁻

¹)Ditch 10 (dammed)


Figure 3.5. (continued) Table 3.4 Proportion of time overland flow (OLF) was recorded at the weir boxes (flow ≥ 0.1 mL s-1) and the difference in proportion of time OLF occurred compared to weir box 7.

Ditch % OLF time % difference to control weir box 7

2012/13 2013/14 2014/15 2012/13 2013/14 2014/15

1 (reprofiled) 66 77 79 4 23 23

2 (control – open) 80 78 83 18 24 27

3 (reprofiled) 40 51 61 -22 -3 5

4 (dammed) 44 71 86 -18 17 30

5 (dammed) 69 94 100* 7 40 44

6 (control – open) 72 56 56* 10 2 0

7 (control – open) 62 54 56

8 (reprofiled) 73 53 72 11 -1 16

9 (control –open) 72 30 32 10 -24 -24

10 (dammed) 98 63 61 36 9 5

11 (reprofiled) 76 63 48 14 9 -8

12 (control – open) 24 25 9* -38 -29 -47

*weir box 5 until May 2014, weir box 6 until July 2014 and weir box 12 until September 2014 only.

of 37

Table 3.5 provides the overall overland flow fluxes for each weir box by year, with corresponding aerial-weighted flow calculations (taking the area that drains to each weir box before ditch blocking in one scenario and the area that drains to each weir box assuming blocked drains are infilled and overflow in the second scenario). The table suggests that the relative contributing area for overland flow has almost tripled between the first year after blocking and the most recent period. The large values of overland flow in catchments 6 and 7 combined suggest that even where the ditches are open, overland flow (and shallow throughflow to 3 cm depth) may be the dominant flow path for water at the study site. Note that the catchment areas for the ditch weirs and the overland flow weir boxes are different and so it is not possible to simply partition the total ditch flow recorded in a ditch weir into overland flow and subsurface flow based on the overland flow recorded at the overland flow weir box. In addition, the overland flow soffit boards extend across the slope capturing flow on the areas of peat between ditches and it may be that, close to ditches, there is less overland flow than midway between ditches. However, on an area-weighted basis across the site, if there is a large proportion of mm of runoff for overland flow in comparison to rainfall, this would suggest that overland flow is dominant across the site. The time series for monthly discharge (m3) per unit rainfall are shown for each overland flow weir box in Figure 3.6. The winter peak in 2013/14 for some weir boxes may reflect snow and ice-melt effects but many of the ditches produced more overland flow per rainfall unit in winter 2013/14 than they did in winter 2012/13 including some of the control ditches, with these high rates being maintained into winter 2014/15. The relationship between overland flow and ditch flow was fairly stable during the study period for control catchments 6 and 7 (Figure 3.7; only data for Ditch 7 are shown because patterns were similar at Ditch 6, and data are only available until July 2014 for overland flow for Ditch 6). The relationship was also stable for most other ditches suggesting that the balance of overland flow and subsurface flow partitioning did not change in the years after blocking. However, for Ditch 4 there was a tendency towards higher overland flow rates in 2014/15 compared to earlier years for comparable ditch flows. The opposite was the case for Ditch 8. For Ditch 5 the relationship between overland flow and ditch flow was similar between years but there was a tendency for higher rates of both in 2013/14 and 2014/15 (note data only available to May 2014), compared with 2012/13 (Figure 3.7).

of 37

Table 3.5. Total annual water fluxes from the overland flow (OLF) weir boxes.

*2011/12 from 1st July 2011 to end of February 2012 aTotal area-averaged OLF across the site (i.e., not a sum of the values in the column above) bThere was a tendency for missing data to be distributed throughout the year and there was no fixed seasonal pattern in rainfall (see Figure 3.6). Therefore the

uplift was calculated using 1 + for each weir box where is the proportion of missing data for that weir box. One shade lighter indicates 5-10% data series missing Lightest shading indicates 10-30% data series missing White cell indicates >30% data series missing

m3 of water

mm of runoff assuming original catchment area

mm of runoff assuming modified catchment area % missing data in time series

Year 2011/12* 2012/13 2013/14 2014/15 2011/12* 2012/13 2013/14 2014/15 2011/12* 2012/13 2013/14 2014/15 2011/12* 2012/13 2013/14 2014/15

Rainfall 1716 2409 1786 1888 1716 2409 1786 1888 0.0 0.0 0.0 0.0

Ditch 1 (reprofiled) 0 141 567 708 0 1406 5671 7075 0 204 822 10254 0.0 2.1 0.2 1.4

Ditch 2 (control – open) 455 504 1260 3213 1987 2200 5504 14030 141 156 391 44 0.0 0.0 0.2 0.5


Ditch 4 (dammed) 2 9 10 51 33 146 172 830 0 2 2 2 0.0 0.0 0.2 0.5

Ditch 5 (dammed) 0 102 504 225 0 2363 11718 5235 0 6 31 3 54.2 0.1 0.2 75.6




Ditch 9 (control –open) 1 46 40 23 23 839 727 421 5 185 160 17 43.0 0.0 5.9 1.2

Ditch 10 (dammed) 146 618 57 25 908 3840 355 156 14 600 55 2 69.7 0.0 0.1 1.2


Ditch 12 (dammed) 52 19 14 7 485 179 126 67 142 52 37 2 43.0 44.0 0.3 46.8

Total 1650 2820 3776 5269 1089a 1862

a 2493

a 3478

a 372

a 636

a 852

a 1188

a 21.1 5.6 2.0 15.8

Uplift for missing datab 1841 2922 3840 5600 1215 1929 2535 3696 415 659 866 1263

Overall OLF capture % of rainfall 71 80 142 196 24 27 48 67

of 37

Mar-14Mar-13Mar-12

400

300

200

100

0

Pre

cip

ita

tio

n (

mm

)

Mar-14Mar-13Mar-12

3

2

1

0

Dis

cha

rge

pe

r ra

infa

ll u

nit

(m

³ m

m⁻¹

)

Ditch 2


Mar-14Mar-13Mar-12

1.2

0.9

0.6

0.3

0.0

Dis

cha

rge

pe

r ra

infa

ll u

nit

(m

³ m

m⁻¹

)

Ditch 4


Mar-14Mar-13Mar-12

1.2

0.9

0.6

0.3

0.0

Dis

cha

rge

pe

r ra

infa

ll u

nit

(m

³ m

m⁻¹

)

Ditch 1


Precipi tation Open

Dammed Reprofi led

Figure 3.6. Monthly discharge (m3) per unit of rainfall (mm) for each overland flow weir box. Note different y-axes scales.

Figure 3.7. Example scatterplots of overland flow and ditch flow for Ditches 4, 5, 7 and 8. Data shown are square root discharges.

of 37

3.3 Water tables Water tables at the study site tended to be very shallow (Table 3.6). Time weighting accounts for variations in the intervals between manual water-table measurements when calculating a mean, thus removing biases that may be caused by a higher frequency of reading at one time of the year compared to another. Time-weighted means are shown in Table 3.6 along with raw means, medians and the interquartile range. At a distance of only 2 m from open ditches, before blocking occurred the mean time weighted water-table depth for each dipwell ranged from 1.7 cm to 20.2 cm (Table 3.6). Of the 24 dipwells located 2 m from ditch edges, eight of these had time weighted water-table depths within 5 cm of the surface for the 6 month period before ditch blocking. Water tables around Ditch 9 (open) tended to be deepest both before and after management interventions across the rest of the site. Using time weighted means, 11 out of 16 dipwells around dammed or reprofiled ditches indicated shallower water tables after management intervention on site compared with the relative conditions 2 m either side of control Ditch 6 (Table 3.7). At some locations the mean net water-table rewetting effect was around 10 cm (e.g., 1.2W (reprofiled), 3.2E (reprofiled), 5.2W (dammed), 10.2W (dammed); Table 3.7). The average time-weighted relative rewetting effect across all of the 2 m dipwells when comparing all blocked and reprofiled ditches with ditch 6 was 0.4 cm, 1.4 cm, 1.8 cm and 1.5 cm in the four study periods (2011/12, 2012/13, 2013/14, 2014/15) after blocking respectively. There was very little difference in the average interquartile range of water-table depth for the 2 m dipwells before and after blocking when comparing the blocked and open ditches. Using time-weighted water-table data, a repeated measures 1-way analysis of variance (ANOVA) was used to test for treatment and time (year) effects for the 2 m dipwells to the west of the ditches as these are where effects of treatment are likely to be greatest due to the site’s gradient. This indicated no significant effect of drainage treatment (open, dammed, reprofiled), but a significant effect of year on water-table depths, with 2013/14 having significantly deeper water tables (by ~1.7 cm) than 2012/13. The lack of a treatment effect might seem somewhat surprising given that the mean water table for the 2 m dipwells next to the open ditches was consistently (i.e., across all four years) deeper than for the wells next to the blocked ditches but is probably a reflection of the relatively small sample size (12 wells in total and four per treatment) and a relatively high variability among the wells adjacent to the open ditches. A separate repeated-measures ANOVA was used to compare the CS 1 m and 3 m dipwells and there was no significant effect of treatment (p = 0.067) from open (average depth: 10.0 cm), dammed (7.2 cm) or reprofiled (5.9 cm) ditches. The results also showed that there was a significant effect of the year after blocking (p= 0.001), with 2011/12 (9.2 cm) > 2014/15 (8.3 cm) > 2013/14 (7.1 cm) > 2012/13 (6.1 cm). There was no effect of distance from the ditch or an interaction effect between year after blocking, treatment and distance. The water-table depths shown in Table 3.6 are relative to the peat surface. However, it is also possible to evaluate the water tables for each dipwell transect relative to a local arbitrary datum (separate datum for each dipwell transect) so that water-table height above that datum can be plotted and compared (Figure 3.8). In all cases the water-table heights above datum, for midpoints between ditches, are much greater than those around the ditches. Such an effect is most likely due to the peat surface being typically higher at the midpoints than adjacent to or within each ditch (e.g., the median peat surface height difference was 28 cm between midpoint dipwells and the dipwells 2 m east of the ditch). The time series in Figure 3.8 also indicate that in most cases the absolute water-table height for the 2 m dipwells that are east and west of each ditch are very similar (except for at Ditch 9 – open), although for Ditch 7 (open) the water-table height on the eastern side was generally a few cm higher than on the western side, while the reverse was true for Ditch 11 (reprofiled). In most ditches for the 2 m dipwell transects the water height above datum in the ditch itself was lower than that in the peat 2 m away. The exceptions were Ditches 6, 8 and 12 where water levels in the ditch appeared to be higher in absolute terms than in the surrounding peat. Ditch 5 showed the closest correspondence of absolute water-table level in the ditch and the surrounding peat. The additional water-table data associated with the gas-flux collars are presented as depth to water table from the peat surface using examples in Figure 3.9 and summarised in Table 3.8. They were also examined with reference to a local datum for each transect (selected examples shown in Figure 3.10). Open ditches 2,

of 37

6, 7, and 9 had much shallower water tables (as depth from peat surface) in the ditches (which would be expected) than 1 or 3 m away on the main peat mass for the collar transects. There was, however, a relatively close correspondence of ditch water-table depth and water-table depth in the surrounding peat at the collar transects for the blocked ditches except at Ditch 8. Unlike for the blocked ditch transects described above, for the collar transects, water-table height above datum was always greater in blocked ditches than in the surrounding peat in the case of Ditches 1, 3 and 11 (all reprofiled ditches), which was not the case at other ditches. Three of the open ditches had absolute water-table height that was far lower than in the surrounding peat. For the collar transects, in comparison to control Ditch 6, water tables at blocked ditches tended to be shallower both 1 m and 3 m from the ditch edge in most cases (Figure 3.11) although during dry periods water tables 1 m from some blocked ditches could fall much further than the water table at Ditch 6 showing variability in the peat’s susceptibility to dry conditions. For the automated dipwells located midway between ditches the records showed a relatively small range in means between dipwells for any given year (e.g., for 2011/12 4.0 cm (Ditch 2) to 12.7 cm (Ditch 7)) (Table 3.9) although these small differences could be important for carbon cycling processes. There were many missing data in the time series for midway dipwell 6, so midway dipwell 7 was used as a control and had the added advantage that it was midway between two open ditches. While there are no data available for these automated dipwells from the period before ditch blocking on site, it is possible to examine the difference between water tables at midpoints between ditches across the site and midway dipwell 7 over time since blocking (Figure 3.12). During the second half of the study period there is an apparent relative drift in water-table depths towards the drier conditions at midway dipwell 7 (between Ditch 6 and 7) for dipwells 4, 5, 10, 11, and 12 (all dammed or reprofiled).

of 37

Table 3.6. Water-table depth (cm) statistics for each transect based on data from manual sampling visits (or extracted for the same time as the manual sampling from automatic records midway and in ditch). Mn = arithmetic mean, Med= median, WM =weighted mean, IQR = inter-quartile range. Empty cells indicate no data.

Midway between ditches 2m east of ditch In ditch 2 m west of ditch Ditch Period Mn Med

. WM IQR Mn Med WM IQR Mn Med WM IQR Mn Med WM IQR

1 rep.

before 4.3 3.7 4.4 2.8 12.6 11.1 15.7 6.8

2011/12 7.5 2.8 9.0 12.8 10.1 8.6 8.9 7.0 2.9 1.5 2.2 7.0 8.1 5.7 7.0 6.8

2012/13 2.1 2.0 2.2 1.0 8.8 5.7 8.8 5.7 0.3 -0.9 0.2 2.1 7.6 5.4 7.3 5.2

2013/14 8.2 2.7 5.9 15.0 9.7 5.4 8.5 12.0 3.9 1.5 2.4 6.4 8.7 5.7 7.8 8.1

2014/15 4.4 -0.8 2.2 14.7 9.7 6.0 8.1 10.4 0.7 -1.2 -0.1 5.2 10.0 7.4 8.8 8.1

2 open

before 6.4 5.8 5.6 2.4 4.9 4.2 4.0 3.4

2011/12 6.2 4.1 5.0 8.8 7.7 5.6 6.7 6.2 3.3 3.1 2.3 6.9 7.8 5.2 6.8 9.5

2012/13 4.5 2.4 4.6 4.7 6.5 5.3 6.2 8.0 -0.5 -0.7 -1.0 5.2 5.3 4.7 4.8 3.6

2013/14 8.9 6.8 6.9 9.5 7.9 5.8 7.0 7.9 0.1 -1.0 -1.8 6.3 6.7 3.5 5.6 8.4

2014/15 5.5 4.0 4.1 6.1 9.4 7.2 7.6 11.1 -4.3 -5.5 -5.2 5.0 7.4 3.9 5.8 11.6

3 rep.

before 11.4 11.1 10.8 2.2 5.9 6.4 5.1 2.3

2011/12 11.3 10.4 10.2 6.9 2.9 1.5 2.3 2.8 -2.0 -2.5 -2.6 5.3 7.2 5.2 6.1 8.1

2012/13 8.0 6.9 7.4 4.9 2.2 0.3 1.8 5.9 -4.7 -5.2 -4.6 1.2 5.9 4.0 5.6 7.2

2013/14 9.3 6.1 7.4 10.5 2.6 -0.4 1.7 5.9 -3.0 -3.6 -3.9 4.7 8.0 4.5 7.1 8.5

2014/15 10.3 10.2 15.2 13.1 4.0 0.8 2.5 7.6 -7.4 -5.4 -6.9 3.1 8.2 4.9 6.6 8.8

4 dam

before 4.6 4.4 4.1 1.7 5.0 3.6 4.8 2.6

2011/12 8.4 4.9 6.8 13.5 7.6 4.7 6.3 7.4 11.5 11.4 10.6 5.9 7.8 6.2 6.6 8.6

2012/13 6.4 3.1 6.2 9.1 6.3 3.2 6.3 6.0 10.2 8.6 9.6 1.8 8.1 6.1 8.0 8.5

2013/14 9.5 7.9 8.0 14.0 8.5 4.8 7.4 11.0 11.3 8.5 9.8 8.9 10.6 5.2 9.0 18.2

2014/15 11.3 7.8 9.0 19.6 9.4 5.0 7.7 12.2 7.1 6.9 5.6 9.5 10.6 5.6 8.6 14.7

5 dam

before 7.1 6.6 6.4 3.5 12.4 12.2 12.0 6.0

2011/12 14.0 12.6 12.4 10.0 9.4 7.9 8.0 8.1 5.9 3.5 4.1 10.0 9.1 4.5 7.0 12.9

2012/13 11.6 9.3 10.9 2.9 8.1 6.2 7.5 8.7 1.7 0.7 3.3 4.1 4.3 1.6 3.8 5.4

2013/14 14.3 10.5 12.2 11.2 8.1 5.0 7.1 8.6 7.3 2.1 5.0 11.3 7.0 2.4 5.9 11.8

2014/15 14.1 12.2 13.8 8.5 9.3 6.1 8.2 9.1 2.9 -0.1 1.3 8.2 7.6 2.7 5.2 13.4

6 open

before 10.1 9.3 9.6 8.9 11.7 13.0 12.0 6.5

2011/12 9.6 7.7 8.9 13.9 12.4 10.8 10.7 12.2 -14.5 -15.3 -15.6 10.6 15.9 14.5 14.9 7.1

2012/13 1.7 2.2 0.9 2.3 10.3 8.4 10.6 9.4 -15.4 -16.6 -15.3 3.2 15.7 12.7 15.9 9.3

2013/14 9.1 6.7 9.9 10.4 13.1 11.1 11.5 15.2 -12.8 -14.7 -14.6 8.5 18.5 14.8 17.2 14.6

2014/15 10.5 12.1 7.2 2.8 14.1 12.2 11.6 14.9 -14.1 -15.5 -15.0 5.4 17.8 14.2 16.6 12.2

7 open

before 3.8 4.2 1.9 3.7 4.3 2.7 2.8 6.1

2011/12 15.1 10.8 13.4 13.4 4.0 3.6 2.9 4.5 9.2 6.4 6.8 13.3 14.7 11.3 13.0 9.8

2012/13 13.1 9.8 12.4 7.9 1.9 1.0 1.7 2.5 0.8 0.2 1.6 1.9 13.6 11.6 13.3 4.8

2013/14 15.5 12.4 12.6 18.8 2.9 2.5 1.0 6.9 4.8 1.2 4.1 5.7 14.5 10.7 13.1 13.8

2014/15 13.9 10.8 12.5 15.8 2.9 1.1 2.3 3.5 6.1 4.4 5.1 15.3 14.6 10.9 12.5 12.7

8 rep.

before 4.2 3.0 3.4 4.2 9.3 8.5 8.7 3.6

2011/12 13.1 10.8 11.2 11.7 9.2 7.5 7.8 6.7 21.7 20.9 21.0 4.0 14.9 14.3 14.0 6.1

2012/13 10.2 8.4 9.4 8.1 7.9 5.8 7.7 5.4 19.3 18.7 18.3 2.4 13.3 11.9 13.1 8.7

2013/14 15.7 14.6 13.2 14.5 9.7 6.5 8.7 10.6 20.9 20.3 20.7 4.4 14.9 12.5 13.6 11.5

2014/15 14.4 12.7 19.4 8.5 10.2 5.8 8.3 11.1 18.9 17.5 18.7 4.9 14.9 12.1 13.3 8.1

9 open

before 20.7 20.7 20.2 4.2 15.0 14.3 14.5 11.1

2011/12 7.0 4.7 5.9 7.5 25.1 24.7 24.0 5.8 22.3 19.9 19.8 17.3

2012/13 7.0 5.4 7.1 7.0 23.8 22.7 23.1 10.3 16.7 13.7 16.6 14.1

2013/14 8.6 5.3 6.8 11.1 23.8 21.6 22.4 13.1 19.4 14.6 17.1 25.0

2014/15 7.3 4.6 5.8 8.0 25.1 20.8 23.0 12.0 22.2 21.2 19.9 17.8

10 dam

before 6.7 5.9 6.5 3.9 12.7 12.1 12.0 4.8

2011/12 12.6 10.9 11.1 8.8 11.7 9.3 10.1 9.9 11.2 11.0 10.5 6.5 8.5 8.5 7.4 7.3

2012/13 10.5 8.3 8.9 7.7 10.8 9.4 10.3 7.0 10.1 9.3 10.3 5.8 7.2 5.3 7.1 7.6

2013/14 16.4 15.5 15.5 17.0 12.2 8.2 10.7 14.5 13.7 13.3 13.4 3.6 10.2 7.9 9.2 7.5

2014/15 11.7 10.0 15.5 5.9 13.1 7.4 10.2 17.3 15.3 15.1 15.7 3.2 9.2 6.6 7.8 8.5

11 rep.

before 7.8 7.7 7.3 3.4 9.7 8.1 8.3 9.8

2011/12 9.9 9.4 7.2 13.1 14.2 13.3 13.7 3.9 -2.9 -3.7 -4.0 7.2 6.9 4.6 5.6 6.5

2012/13 -1.1 -1.5 -1.6 2.6 14.5 14.3 14.3 5.1 -6.7 -6.9 -6.9 8.8 5.4 3.6 5.0 4.7

2013/14 6.6 3.3 6.5 14.6 15.2 12.9 14.6 6.9 -7.2 -7.8 -8.1 3.5 7.1 3.7 6.3 6.2

2014/15 6.8 3.7 6.9 10.2 16.6 12.9 15.0 9.8 -8.7 -9.8 -9.5 8.8 6.6 3.2 5.0 7.0

12 dam

before 7.4 6.0 6.6 4.0 2.4 1.7 1.7 2.7

2011/12 3.9 1.9 2.3 9.5 8.4 5.9 6.9 8.9 -8.8 -7.9 -9.9 9.5 13.5 12.9 11.8 12.0

2012/13 1.0 -1.5 1.1 6.5 6.6 4.3 6.6 7.0 -9.9 -12.5 -10.3 6.8 11.7 9.7 11.7 7.3

2013/14 6.6 3.0 4.4 12.8 7.9 4.6 7.0 11.9 -7.1 -7.3 -7.5 8.1 14.5 11.4 13.4 12.2

2014/155

3.6 -0.2 2.2 10.6 8.6 4.8 6.2 13.0 -8.0 -8.9 -8.1 5.9 15.1 11.0 12.7 15.5

of 37

Table 3.7. Mean time-weighted water-table depth offset to control dipwells at ditch 6 after blocking minus mean offset to control dipwells before blocking (cm). A positive value indicates a relative wetting effect compared to the control.

Treatment Dipwell 2011/12 2012/13 2013/14 2014/15

Reprofiled 1.2E -3.4 -3.4 -2.2 -1.7

1.2W 11.5 12.2 13.0 11.5

Open 2.2E 0.0 0.4 0.5 0.1

2.2W 0.1 3.2 3.6 2.8

Reprofiled 3.2E 9.6 10.0 11.0 10.3

3.2W 1.9 3.4 3.2 3.1

Dammed 4.2E -1.1 -1.2 -1.5 -1.6

4.2W 1.1 0.7 1.0 0.8

Dammed 5.2E -0.5 0.0 1.2 0.2

5.2W 7.9 12.1 11.3 11.5

Open 7.2E 0.1 1.2 2.8 1.6

7.2W -7.2 -6.6 -5.0 -5.1

Reprofiled 8.2E -3.3 -3.3 -3.4 -2.9

8.2W -2.4 -0.5 0.3 0.1

Open 9.2E -2.7 -1.9 -0.3 -0.8

9.2W -2.5 1.8 2.5 -0.8

Dammed 10.2E -2.6 -2.9 -2.3 -1.8

10.2W 7.5 8.8 7.9 8.7

Reprofiled 11.2E -5.3 -6.0 -5.5 -5.8

11.2W 5.6 7.2 7.2 8.0

Dammed 12.2E 0.7 1.0 1.5 2.3

12.2W -7.2 -6.1 -6.5 -6.4

of 37

1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

2 m east

in ditch

2 m west

midway

1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

Ditch 1 (reprofiled) Ditch 2 (open)


1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

Ditch 5 (dammed) Ditch 6 (open)

Ditch 7 (open) Ditch 8 (reprofiled)

Figure 3.8. Water-table height above local datum for each ditch transect (continued on next page).

of 37

1/1/151/3/151/3/11

300

200

100

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

2 m east

in ditch

2 m west

midway

1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

1/1/151/3/151/3/11

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

1/1/151/3/151/3/11

400

350

300

250

He

igh

t a

bo

ve

loca

l da

tum

(cm

)

Ditch 9 (open) Ditch 10 (dammed)


Figure 3.8. (continued)

1/1/151/3/131/3/11

0

20

40

0 m

1 m

3 m

1/1/151/3/131/3/11

0

20

40

Wa

ter-

tab

le d

ep

th (

cm)

1/1/151/3/131/3/11

0

20

40

Ditch 1 (reprofiled)

Ditch 4 (dammed)

Ditch 6 (open)

Figure 3.9. Time series of water-table depth for the gas collars at 0 m, 1 m and 3 m from the ditch for three example ditches.

of 37

1/1/151/3/131/3/11

420

400

380

0 m1 m3 m

1/1/151/3/131/3/11

360

340

320

He

igh

t a

bo

ve

lo

cal

da

tum

(cm

)

1/1/151/3/131/3/11

400

350

300


Ditch 4 (dammed)

Ditch 6 (open)

Figure 3.10. Time series of water-table height above local datum for the gas flux collars at 0 m, 1 m and 3 m from the ditch for three example ditches. Datum is specific to each ditch.

of 37

Table 3.8. Depth to water table (cm) summary data for gas-flux collars, July 2011 to August 2014 (0 is in ditch, 1 is 1 m from ditch edge, 3 is 3 m from ditch edge). IQR = interquartile range.

Treatment Ditch/Dipwell Mean Minimum Median Maximum IQR

Reprofiled 1.0 7.3 0.7 6.0 25.8 6.4

1.1 6.2 0.0 5.8 18.4 4.7

1.3 4.1 -1.9 3.4 15.3 6.5

Open 2.0 -3.2 -8.0 -2.9 3.7 4.7

2.1 3.0 -2.1 1.5 15.5 5.5

2.3 5.7 -1.3 4.5 19.1 6.2

Reprofiled 3.0 11.1 2.6 9.3 28.6 7.9

3.1 6.5 -0.3 4.8 23.1 4.9

3.3 5.3 -2.9 4.8 15.5 5.6

Dammed 4.0 5.7 -0.8 3.7 26.5 7.8

4.1 9.1 2.1 7.1 25.2 8.0

4.3 2.8 -1.4 2.3 9.4 3.5

Dammed 5.1 4.4 -2.1 1.5 31.2 5.2

5.3 8.4 1.9 7.4 19.3 7.8

Open 6.0 9.9 3.4 7.8 27.0 5.1

6.1 -1.7 -12.7 -2.0 13.9 5.1

6.3 10.9 4.9 9.8 45.2 5.3

Open 7.0 10.3 2.5 8.4 36.2 5.9

7.1 -9.2 -14.2 -9.2 -3.2 2.6

7.3 6.5 0.0 5.4 38.9 5.6

Reprofiled 8.0 11.9 6.1 10.8 41.7 4.2

8.1 17.7 7.7 18.1 31.0 8.9

8.3 8.9 1.1 7.0 26.2 7.1

Open 9.0 6.9 -2.5 5.2 17.9 7.3

9.1 4.5 -9.5 4.2 19.6 4.4

9.3 11.4 4.2 11.0 44.8 6.3

Dammed 10.0 11.1 3.5 9.6 41.0 4.9

10.1 3.5 -3.0 3.2 13.0 5.4

10.3 9.8 6.3 9.1 21.4 2.7

Reprofiled 11.0 8.7 2.5 7.0 20.0 7.3

11.1 3.9 -2.4 2.3 18.8 6.4

11.3 8.6 -0.5 7.9 26.6 10.2

Dammed 12.0 5.0 -1.1 2.5 22.3 5.4

12.1 4.5 -0.2 3.8 15.9 5.6

12.3 8.1 3.2 7.4 20.7 3.9

of 37

Figure 3.11. Difference between water-table depth for collar dipwells at Ditch 6 and collar dipwells at other ditches over time. Positive values indicate shallower (closer to surface) water tables near the study ditch compared to Ditch 6.

of 37

Table 3.9. Depth to water table (cm) summary data from automated logger records at the mid-point on the peatland between drains. IQR = interquartile range.

Ditch Year % missing Mean Median Maximum IQR

1 rep. 2011/12 24.0 5.2 2.5 33.1 3.6

2012/13 9.9 2.8 1.9 23.7 1.9

2013/14 6.4 5.0 1.3 39.0 7.3

2014/15 12.3 3.8 -0.8 30.6 8.4

2 open 2011/12 2.9 4.0 2.9 19.4 4.6

2012/13 1.1 3.2 2.4 17.0 3.5

2013/14 6.3 6.3 5.1 26.8 6.2

2014/15 1.8 4.1 2.5 20.6 5.9

3 rep. 2011/12 0.0 9.4 8.8 28.6 3.8

2012/13 0.3 5.7 5.4 22.6 4.4

2013/14 6.4 6.9 5.2 33.3 5.1

2014/15 58.1 8.9 5.9 29.1 7.2

4 dam 2011/12 9.4 6.5 3.9 26.7 8.1

2012/13 1.1 4.3 2.2 22.2 5.3

2013/14 1.5 6.9 4.1 30.6 11.3

2014/15 42.3 11.9 10.3 32.5 15.2

5 dam 2011/12 0.3 11.1 9.9 39.2 5.4

2012/13 0.6 9.2 8.5 31.2 3.5

2013/14 6.3 11.8 9.4 44.5 5.1

2014/15 11.1 13.0 11.0 37.7 4.9

6 open 2011/12 35.9 5.9 3.2 36.4 9.2

2012/13 32.0 3.7 3.2 16.3 3.6

2013/14 3.5 8.6 6.5 43.6 6.2

2014/15 53.1 14.2 10.9 53.8 13.7

7 open 2011/12 0.2 12.7 9.6 36.0 6.6

2012/13 0.8 10.2 8.2 30.8 3.1

2013/14 6.4 11.7 7.5 39.6 11.5

2014/15 0.1 12.2 9.3 40.7 10.8

8 rep. 2011/12 1.1 10.5 9.0 37.2 7.1

2012/13 2.1 7.1 6.0 29.2 6.2

2013/14 10.6 12.5 10.8 42.5 10.7

2014/15 1.2 13.8 11.3 41.6 8.7

9 open 2011/12 0.7 5.0 3.7 26.9 3.5

2012/13 0.6 5.5 5.1 18.7 2.7

2013/14 6.3 6.3 4.3 30.5 4.5

2014 2.1 6.2 4.1 26.7 5.3

10 dam 2011/12 0.3 10.3 9.1 33.2 4.8

2012/13 7.2 7.6 7.2 27.6 3.7

2013/14 38.2 7.5 5.0 45.2 7.3

2014/15 1.9 11.4 9.5 34.8 6.3

11 rep. 2011/12 12.0 7.0 5.1 32.7 10.9

2012/13 42.6 1.2 0.0 17.8 3.8

2013/14 8.3 6.6 3.9 40.9 9.1

2014/15 7.8 8.5 3.6 72.3 11.6

12 dam 2011/12 19.8 2.7 0.7 24.1 5.6

2012/13 30.5 0.6 -0.6 18.2 3.1

2013/14 12.3 4.1 1.8 32.9 7.4

2014/15 16.6 3.9 0.9 27.2 6.7

of 37

Jan-15Mar-13Mar-11

10

5

0

-5Mean o

ffse

t to

mid

7 d

ipw

ell (

cm)


Ditch 2 (open)


Ditch 4 (dammed)

Jan-15Mar-13Mar-11

10

5

0

-5Mean o

ffse

t to

mid

7 d

ipw

ell (

cm)

Ditch 5 (dammed)

Ditch 8 (reprofiled

Ditch 9 (open)

Jan-15Mar-13Mar-11

10

5

0

-5Me

an o

ffse

t to

mid

7 d

ipw

ell

(cm

)

Ditch 10 (dammed)


Ditch 12 (dammed)

Figure 3.12. Monthly mean offset (based on 2-hourly data time series) between study midpoint dipwell and midpoint 7 dipwell (the control). A positive value indicates shallower water-table conditions compared to the control.

4. Discussion

4.1 Catchment area assessment The 82 % runoff efficiency (proportion of rainfall converted into discharge) for the 2 ha hillslope before ditch blocking took place is in line with previous water budgets for headwater peatlands (Evans et al., 1999; Holden, 2006; Holden et al., 2012). For the period before ditch blocking we can be most confident about the cumulative catchment area for the twelve ditch weirs. However, even during this first period of the study there are two ditches (4 and 9) that appear to produce far more aerially-weighted discharge than rainfall. Such data provide clear evidence that the water (and carbon) source areas for these ditches are different from those defined by the surface topography alone. It may be that subsurface springs, pipes and other throughflow pathways result in source areas for those ditches which stretch beyond the topographically-defined catchment. Indeed, water chemistry data for Ditch 4 suggests that there may be a mineral groundwater source for that ditch (see Evans et al. (2016); Appendix E). Hence, at scales of around 1000 to 3000 m2 which are typical surface catchment areas for the outlets of peatland ditches, caution must be taken when calculating water budgets and it may be necessary to reconsider the findings from earlier studies that have looked at aerially-weighted flow rates and aquatic carbon fluxes at such scales, including those from ditch and ditch-blocking studies. Fortunately, at a one order of magnitude greater scale (20000 m2), such effects do not appear to occur. However, it is still possible that subsurface sources for the monitored part of the hillslope occur outside this cumulative topographic area, but logically such effects should decrease as catchment area scale increases.

of 37

4.2 Slow change over time There is strong evidence of both a step change at the study site as a result of ditch blocking and a gradual change over time after ditch blocking. The ditch blocking had an immediate effect on ditch flows, with an approximately five-fold reduction in flow from the ditches that were blocked. After ditch blocking there was a gradual overall increase in runoff from the site over time so that for each unit of rainfall the site produced a greater volume of water which gradually (but significantly) increased over time in the years after blocking. Such a change was mainly manifest as an increase in baseflow from the ditches, with a reduction in completely dry flow periods, and more gently sloping flow duration curves. Thus the peat system showed evidence of a lagged hydrological response. Such lagged responses have been shown for water-table records before (slow recovery in water-table depths and slow reduction in water-table variability (Holden et al., 2011; Wilson et al., 2010)) but never for water flows in a blanket peatland channel system. It is not clear why the flows should increase over time in the years after blocking, compared to the year immediately after blocking, but it is quite possible that leaks slowly developed in the dam network. The re-packed peat that formed the dams may not be stable and could be prone to piping and cracking caused by subsidence or the high seepage force associated with the large hydraulic gradient between the upper and lower part of the dam. It may also be that adjustments to surface and subsurface flowpaths occurred such as new routes for water to bypass dams and flow back into ditches around vegetation on the peat surface, or changes in subsurface pipe connectivity associated with ponding in ditches. When using area-weighted discharge based on the original surface-derived catchment areas for each drain, the total discharge efficiency was found to be greater than 100% for several drains. Thus their real catchment areas must have increased over time due to the ditch blocking activity and spillage of water from one ditch to another. However, the whole system had not shifted to behave as if the blocked drains had completely infilled because, when the cumulative catchment area for the weirs was used in the infilling scenario for the overall study site, runoff efficiency was > 100%. The figure is so high because the catchment area in the infilling scenario is much smaller than for the open ditch scenario. The system appears to be operating in the latter part of the study, in terms of catchment source areas, somewhere in between that shown in Figure 2.1 and that shown in Figure 2.2. Overland flow was only monitored for the period after blocking. There were very long periods of saturation-excess overland flow production on the site, particularly around blocked drains. Overland flow continued to occur on the slopes near all blocked drains for more than 50 % of the time after blocking. The relationship between overland flow and ditch flow was stable from year to year for the open control ditches. However for some (but not all) of the blocked ditches the relationship shifted from year to year suggesting that long-term changes to the hydrological system as a result of ditch blocking were spatially variable, with lagged effects in some areas and for some processes. Evidence from some of the water-table records also suggests lag effects such as the deepening of the water-tables at the mid-point between ditches in comparison to the control in the latter part of the record. This may either be a recovery effect from site disturbance operations and machinery, or it may be further evidence to suggest that the site became ‘leaky’ and that initial successful rewetting of inter-ditch areas was reduced as ditch dams (and the ditch-flow weirs) started to release more water in the latter half of the study.

4.3 Water-table change Overall, water-table depths on site were relatively shallow, similar to what one would expect to see on an intact and fully functioning blanket bog (Evans et al., 1999; Gilman, 1994; Lindsay, 2010). These data show that ditch drainage was not very effective at the site. This is potentially due to high rainfall at the site, low hydraulic conductivity of the peat and the fact that some of the ditches were orientated in an almost downslope direction. Nevertheless, at some locations the mean net time-weighted water-table rewetting effect, compared to open ditch conditions, was around 10 cm while the average was < 2 cm showing high

of 37

spatial variability. Such variability is to be expected given that the ditches were intermittently dammed rather than being completely infilled and hence there will be deep and shallow parts of ditch pools and concentrated locations where ditches overflow onto the surrounding peat surface. However, the findings also suggest that future ditch blocking work on upland blanket peat could be informed by either prior testing of site water tables or by spatial topographic modelling to help determine whether there are sites which could potentially undergo larger water-table improvements through ditch blocking than were observed at our study site. Comparison of some of the water-table height above datum records suggest some steep hydraulic gradients occur within 3 metres of the ditches even after ditch blocking. Thus downslope running ditches probably continue to act as a sink for surrounding peat water in the immediate vicinity of the ditch when the water level in the ditch is below that in the surrounding peat. The volume of subsurface flow that enters the ditches will also depend upon the hydraulic conductivity which was not measured on site. The ditches will additionally have received overland flow and near-surface flow (upper few cm) from inter-ditch areas because (a) they were not perfectly aligned downslope; (b) flows across hillslopes are unlikely to always be perfectly aligned in a downslope direction and there is generally some sideways shedding of water; and (c) the topography will force overland flow into ditches especially if there has been wastage and consolidation of the peat near ditches and the peat surface is domed between ditches. Thus downslope inter-ditch areas will be deprived of hillslope water when ditches are open or still partly function after blocking even if they are predominately orientated in a downslope direction. As would be expected, some blocked ditch water levels were higher in absolute terms than in the surrounding peat and so water would tend to flow from the ditch into the peat at those points. Comparison of the 2 m dipwell transects and the collar transects on the same ditch suggests that water may be drawn into the ditch from the peat at some points and then from the ditch back into the peat at other points along the ditch course.

5. Conclusions The hydrological analysis at the study site has shown that the site is a typical flashy blanket peatland system, dominated by overland flow, but with evidence of subsurface flow connectivity that extends beyond the topographic boundaries of small ditch catchments. There was extremely high variability in flow rates between ditches which had very similar surface catchment areas. Hence caution is needed when upscaling from studies that may have only collected evidence on hydrological flows and aquatic carbon fluxes from one or two ditches (or blocked ditches) (e.g., Armstrong et al., 2010; Gibson et al., 2009). At small individual ditch catchment scales, caution must also be taken when calculating aerially-weighted water and carbon budgets. The evidence suggests that it may be necessary to reconsider the findings from earlier studies that have looked at aerially-weighted flow rates and aquatic carbon fluxes at such scales, including those from ditch and ditch-blocking studies. While ditch blocking had an immediate effect on ditch flows, the analysis shows that there has also been long-term change in the hydrology of the system in the years following ditch blocking. There was some evidence (ditch, overland flow and water-table data) to suggest that the system has become more ‘leaky’ since the initial restoration works were carried out with a greater volume of water per mm of rainfall flowing down the ditch or former ditch channels (but in the form of slow seepage and baseflow, rather than high flow peaks). It is not clear why this has occurred and several lines of investigation should be explored including the possibility that the dams are leaking at an increasing rate, that new flow routes have formed allowing water to enter back into ditches that was previously distributed away from ditches, and that subsurface connectivity of bypassing flow (e.g., pipeflow) may be important on site. Thus the analysis has shown the need for long-term monitoring studies to test whether findings in the initial post-restoration phase still apply several years later and also as part of testing the robustness of management intervention measures in later years after ditch blocking. It may be that, because the predominant orientation of the ditches is downslope, leakiness changes over time are more likely than at sites where ditches run in a more cross-slope direction.

of 37

The hydrological data analysed herein underpin aquatic carbon flux calculations and also gas flux modelling for the site. These parts of the study are presented and discussed in Appendices D and E.

6. References

Acreman, M., and Holden, J. 2013. How wetlands affect floods. Wetlands, 33, 773-786, doi: 10.1007/s13157-013-0473-2.

Armstrong, A., Holden, J., Kay, P., Foulger, M., Gledhill, S., McDonald, A.T., and Walker, A. 2009. Drain-blocking techniques on blanket peat: a framework for best practice. Journal of Environmental Management, 90, 3512-2519, doi:10.1016/j.jenvman.2009.06.003.

Armstrong, A. Holden, J., Kay, P., Francis, B., Foulger, M., Gledhill, S., McDonald, A.T., Walker, A. 2010. The impact of peatland drain-blocking on dissolved organic carbon loss and discolouration of water; results from a national survey. Journal of Hydrology, 381, 112-120.

Baird, A.J., Holden, J., and Chapman, P. 2009. A Literature Review of Evidence on Emissions of Methane in Peatlands. Defra Project SP0574, University of Leeds, Leeds, UK, 54 pp. Produced for UK Government's Department of Environment Fisheries and Rural Affairs and available at: http://randd.defra.gov.uk/Default.aspx?Module=More&Location=None&ProjectID=15992 (last accessed 16.01.2015).

Baldock, D., Hermans, B., Kelly, P., and Mermet, L. 1984. Wetland Drainage in Europe: The Effects of Agricultural Policy in Four EEC Countries. International Institute for Environmental and Development and the Institute for European and Environmental Policy, Nottingham, 166 pp.

Beadle, J., Brown, L.E., and Holden, J.. 2015. Biodiversity and Ecosystem Functioning in Natural Peat Pools and Those Created by Rewetting Schemes. Wiley Interdisciplinary Reviews: Water, 2, 65-84, doi: 10.1002/wat2.1063.

Cunliffe, A., Baird, A.J., and Holden, J. 2013. Hydrological hotspots in peatlands: spatial variability in hydraulic conductivity around natural soil pipes. Water Resources Research, 49, 5342-5354.

Evans, M.G., Burt, T.P., Holden, J., and Adamson, J.K. 1999. Runoff generation and water table fluctuations in blanket peat: evidence from UK data spanning the dry summer of 1995. Journal of Hydrology, 221, 141-160.

Evans, C., Chapman, P., Green, S.M., Baird, A.J., Holden, J., Cooper, D., Peacock, M., Callaghan, N., and Robinson, I. 2016. Appendix E in: Green, S.M., Baird, A.J., Evans, C., Ostle, N., Holden, J., Chapman, P.J., and McNamara, N. Investigation of Peatland Restoration (Grip Blocking) Techniques to Achieve Best Outcomes for Methane and Greenhouse Gas Emissions/Balance: Field Trials and Process Experiments Final Report. Defra Project SP1202, University of Leeds, Leeds.

Gibson, H.S., Worrall, F., Burt, T.P., and Adamson, J.K. 2009. DOC budgets of drained peat catchments: implications for DOC production in peat soils. Hydrological Processes, 23, 1901-1911.

Gilman, K. 1994. Hydrology and Wetland Conservation. John Wiley, Chichester, 101 pp.

Green, F.H.W. 1974. Changes in artificial drainage, fertilizers, and climate in Scotland. Journal of Environmental Management, 2, 107-121.

Green, S.M., Baird, A.J., Evans, C., Peacock, M., Holden, J., Chapman, P., Gauci, V., and Smart, R. 2016. The effect of drainage ditch blocking on the CO2 and CH4 budgets of blanket peatland. Appendix D in: Green, S.M, Baird, A.J., Evans, C., Ostle, N., Holden, J., Chapman, P.J., and McNamara, N. Investigation of Peatland Restoration (Grip Blocking) Techniques to Achieve Best Outcomes for Methane and Greenhouse Gas Emissions/Balance: Field Trials and Process Experiments Final Report. Defra Project SP1202, University of Leeds, Leeds.

of 37

Haapalehto, T.O., Vasander, H., Jauhiainen, S., Tahvanainen, T., and Kotiaho, J.S. 2011. The effects of peatland restoration on water-table depth, elemental concentrations, and vegetation: 10 years of changes. Restoration Ecology, 19, 587-598, doi: 10.1111/j.1526-100X.2010.00704.x.

Holden, J. 2004. Hydrological connectivity of soil pipes determined by ground-penetrating radar tracer detection. Earth Surface Processes and Landforms, 29, 437-442.

Holden, J. 2005. Peatland hydrology and carbon cycling: why small-scale process matters. Philosophical Transactions of the Royal Society A, 363, 2891-2913.

Holden, J., 2006. Peat hydrology. In Martini, I.P., Cortizas, A.M., and Chesworth, W. (eds), Peatlands: Basin Evolution and Depository of Records of Global Environmental and Climatic Changes. Elsevier, Amsterdam, pp. 319-346.

Holden, J., and Burt, T.P. 2002. Piping and pipeflow in a deep peat catchment. Catena, 48, 163-199.

Holden, J., and Burt, T.P. 2003a. Hydraulic conductivity in upland blanket peat: measurement and variability. Hydrological Processes, 17, 1227-1237.

Holden, J., and Burt, T.P. 2003b. Hydrological studies on blanket peat: the significance of the acrotelm-catotelm model. Journal of Ecology, 91, 86-102.

Holden, J., Evans, M.G., Burt, T.P., and Horton, M. 2006. Impact of land drainage on peatland hydrology. Journal of Environmental Quality, 35, 1764-1778, doi: 10.2134/jeq2005.0477.

Holden, J., Shotbolt, L., Bonn, A., Burt, T.P., Chapman, P.J., Dougill, A.J., Fraser, E.D.G., Hubacek, K., Irvine, B., Kirkby, M.J., Reed, M.S., Prell, C., Stagl, S., Stringer, L.C., Turner, A., and Worrall, F. 2007. Environmental change in moorland landscapes. Earth-Science Reviews, 82, 75-100.

Holden, J., Smart, R.P., Dinsmore, K., Baird, A.J, Billett, M.F., and Chapman, P.J. 2012. Natural pipes in blanket peatlands: major point sources for the release of carbon to the aquatic system. Global Change Biology, 18, 3568-3580.

Holden, J., Wallage, Z.E., Lane, S.N., and McDonald, A.T. 2011. Water table dynamics in drained and restored blanket peat. Journal of Hydrology, 402, 103-114.

Jones, J.A.A. 1997. Pipeflow contributing areas and runoff response. Hydrological Processes, 11, 35-41.

Lane, S.N., and Milledge, D.G. 2012. Impacts of upland open drains upon runoff generation: a numerical assessment of catchment-scale impacts. Hydrological Processes, 27, 1701-1726.

Lewis, C., Albertson, J., Xu, X., and Kiely, G. 2011. Spatial variability of hydraulic conductivity and bulk density along a blanket peatland hillslope. Hydrological Processes, 26, 1527-1537, doi: 10.1002/hyp.8252.

Limpens, J., Berendse, F., Blodau, C., Canadell, P., Freeman, C., Holden, J., Roulet, N., Rydin, H., and Schaepman-Strub, G. 2008. Peatlands and the carbon cycle: from local processes to global implications - a synthesis. Biogeosciences, 5, 1475-1491.

Lindsay, R, 2010. Peat Bogs and Carbon: A Critical Synthesis. Royal Society for the Protection of Birds, Edinburgh, 315 pp.

McCarter, C.P.R., and Price, J.S. 2013. The hydrology of the Bois-des-Bel bog peatland restoration: 10 years post-restoration. Ecological Engineering, 55, 73-81, doi: 10.1016/j.ecoleng.2013.02.003.

Newson, M.D. 1992. Conservation management of peatlands and the drainage threat: hydrology, politics and the ecologist in the UK. In Bragg, O.M., Hulme, P.D., Ingram, H.A.P., and Robertson, R.A. (eds), Peatland Ecosystems and Man: An Impact Assessment. University of Dundee, Dundee, pp. 94-103.

Parry, L.E., Holden, J., and Chapman, P.J. 2014. Restoration of blanket peatlands. Journal of Environmental Management, 133, 193-205.

Price, J.S. 1992. Blanket Bog in Newfoundland 2. Hydrological Processes. Journal of Hydrology, 135, 103-119.

of 37

Price, J.S., and Schlotzhauer, S.M. 1999. Importance of shrinkage and compression in determining water storage changes in peat: the case of a mined peatland. Hydrological Processes, 13, 2591-2601.

Smart, R.P., Holden, J., Dinsmore, K., Baird, A.J., Billett, M.F., Chapman, P.J., and Grayson, R. 2013. The dynamics of natural pipe hydrological behaviour in blanket peat. Hydrological Processes, 27, 1523-1534, doi: 10.1002/hyp.9242.

Wilson, L., Wilson, J., Holden, J., Johnstone, I., Armstrong, A., and Morris, M. 2010. Recovery of water tables in Welsh blanket bog after drain-blocking: discharge rates, timescales and the influence of local conditions. Journal of Hydrology, 391, 377-386.

Worrall, F., Armstrong, A., and Holden, J. 2007. Short-term impact of peat drain-blocking on water colour, dissolved organic carbon concentration, and water table depth. Journal of Hydrology, 337, 315-325.

Yu, Z., Loisel, J., Brosseau, D.P., Beilman, D.W., and Hunt, S.J. 2010. Global peatland dynamics since the Last Glacial Maximum. Geophysical Research Letters, 37, L13402, doi:10.1029/2010GL043584.

appendix c. the impact of ditch blocking on blanket

Documents