• infertility;
  • Japanese larch;
  • nutrition;
  • opencast coal;
  • reclamation;
  • revegetation;
  • waterlogging


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

1. Variable growth of Larix leptolepis (Japanese larch) has been observed on restored opencast coal workings in South Wales. This has implications for the restoration of such sites. A study of the relationship of tree growth with minespoil chemical, nutritional, physical and hydrological factors was carried out.

2. Tree growth was positively related to nitrogen and phosphorus foliar concentrations, but negatively to those of magnesium. Seventy-three per cent of the variation in tree growth was explained by variation in foliar chemistry.

3. Soil pH and extractable magnesium were negatively correlated with tree growth, with cation exchange capacity positively related to it.

4. Minespoils had bulk densities that commonly exceeded 1·7 g cm–3 below 0·2 m depth. Stone contents were high and typically 25% by volume.

5. Root systems of trees excavated were characterized by a high root density within 0·3 m of the minespoil surface. Restricted rooting was attributed to high bulk density and the incidence of shallow water tables.

6. Waterlogging during the spring and early summer, and the consequent presence of anaerobic soil conditions during periods of active growth, was found to be detrimental to tree growth.

7. The study suggests that landform design, selection of suitable soil or soil-forming materials, spoil placement technique and appropriate species choice are central to the future success of forestry schemes on restored ground in South Wales.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Many factors affect the successful re-establishment of vegetation on a mining site after mineral extraction has finished (Bradshaw 1983; Moffat & Buckley 1995). Establishing vegetation on spoils produced from coal mining often presents particularly challenging problems. These include compaction, poor water-holding capacity, infertility, high acidity, salinity and hostile temperatures regimes (Moffat & McNeill 1994; Richards, Moorehead & Laing Ltd 1996).

Significant areas of Britain are still affected by both deep mining and opencast mining operations. For example, in 1988 about 4700 ha and 8500 ha, respectively, were affected by these two types of coal extraction in England alone (Department of the Environment 1991a,b). Large areas of commercial forestry plantations established by the Forestry Commission during the 1930s and 1940s in South Wales have been disturbed by opencast mining, and traditionally these have been replanted once mining has finished. Broad (1979) calculated that the total area of the 18 former opencast sites (OCCS) restored to forestry by 1975 was about 820 ha. We estimate that a further eight sites covering an additional 470 ha were restored to forestry between 1975 and 1988, bringing the total area to just under 1300 ha. The greatest concentration of former opencast sites is located in and around the upper Neath valley in the central northern part of the South Wales coalfield.

Between 1942 and 1980, little effort was made to conserve the original soil cover on the majority of sites in the region operated by British Coal (Grimshaw 1992). Any soil that was salvaged was devoted to agricultural restoration, while re-afforestation was based on minespoils composed of crushed overburden strata. These mixtures of shales, mudstones and sandstones have been used as ‘soil-forming’ materials over most of the area restored to forestry.

The growth of the four principal species planted on these sites, Pinus contorta var. latifolia S. Wats. (lodgepole pine), Pinus nigra var. maritima (Ait.) Melville. (Corsican pine), Pinus sylvestris L. (Scots pine) and Larix leptolepis (Siebold & Zuccarini) Gordon. (Japanese larch), has always been characterized by extreme heterogeneity. Trees on some sites are of acceptable size and habit, but the poorest are small and stunted and it is unlikely that they will ever form a commercial crop.

Research to investigate the causes of poor tree growth on restored ground in the late 1950s took the form of species trials (White 1959). The emphasis moved to studies of mineral nutrition in the 1960s (Mayhead, Broad & Marsh 1974), and in the 1970s towards considering cultivation and landform (Broad 1979). Unfortunately this research relied heavily on simple measurements of tree growth rate rather than on quantitative measurements of minespoil properties themselves. In May 1988, the British Coal Opencast Executive appointed the Forestry Commission Research Agency to perform a detailed study of the site factors affecting tree growth on the restored ground in South Wales.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Site selection

Study sites were selected on former opencast coal workings located in and around the upper Neath valley (51°45′N, 3°35′W), where substantial and systematic differences in tree performance occurred. Site conditions were compared at locations where trees demonstrated extremes of performance, both within and across a number of representative sites, to identify important minespoil properties responsible for variation in tree response. Larix leptolepis was chosen for the study because it has demonstrated the greatest commercial potential on restored sites in South Wales.

Pure, or very near pure, stands of L. leptolepis between 5 and 20 years old were examined using Forestry Commission stock maps and field inspection. Ten study sites were selected at the Fyndaff, Dunraven, Bryn Pica and Abercrave OCCS. They ranged in area between 0·07 ha and 0·11 ha (mean 0·08 ha; approximately 150 trees). The age of the trees was between 6 and 19 years in 1988.

Nine of the study sites were subdivided qualitatively into ‘good’, ‘medium’, ‘poor’ and, exceptionally, ‘very poor’ subsites on the basis of tree growth and general condition. Subdivision was not carried out at the tenth site, Fyndaff 1, where tree growth was comparatively uniform. At each study site, measurements were made of altitude, slope angle, aspect, type of ground cultivation, amplitude and wavelength of any ridges and furrows, planting positions, and the presence or absence of watercourses. A summary of the physical characteristics of the 10 study sites is presented in Table 1. Full details of the sites are given in Bending (1993).

Table 1.  Details and measurements made at each study site
Study siteNational grid referencePlanting yearAltitude (m above sea level)Slope (°)AspectFoliar analysisChemical analysis (full)Chemical analysis (reduced)Monitoring of water tablesExcavation of root systems
  1. Full suite of chemical analyses = pH, electrical conductivity, cation exchange capacity, loss-on-ignition, total nitrogen, extractable phosphorus, potassium and magnesium.

  2. Reduced suite = pH, total nitrogen.

Abercrave 1SN 805111196923032SSWYesYes Yes 
Abercrave 2SN 80411219712056SSWYesYes Yes 
Abercrave 3SN 80811619712203NNEYesYes YesYes
Bryn Pica 1SO 008044197332512NWYes Yes  
Bryn Pica 2SO 004042197336018WSWYesYes YesYes
Dunraven 1SN 899053197922524WSWYes Yes  
Dunraven 2SN 89704719802604WYesYes  Yes
Fyndaff 1SN 912045198232015NNEYesYes Yes 
Fyndaff 2SN 908048198229036NWYes YesYes 
Fyndaff 3SN 90405219812407WYesYes YesYes

Tree growth

At each subsite, height was measured on five representative trees in November 1988. These were chosen to be as close to one another as possible, to minimize any effect of minespoil variability within a subsite. Top height/age curves for L. leptolepis were used to provide estimates of general yield class (mean annual increment of a stand in m3 ha–1 year–1) (Hamilton & Christie 1973) and allowed comparisons of growth between the different aged study sites.

Foliar chemistry

The nutrient status of the trees in each subsite was examined by foliar analysis, using standard methods for the collection and preparation of foliage (Everard 1974; Ballard & Carter 1985). Samples were collected in mid-September 1988, from the uppermost lateral shoot, immediately below the leader, on each of the five representative trees within each subsite. Needle samples were oven dried at 70 °C for 24 h on receipt at the Forestry Commission Research Station's laboratory at Alice Holt, Hampshire, UK. Individual ground samples were digested with a mixture of sulphuric acid and hydrogen peroxide. Nitrogen and phosphorus concentrations were measured by colorimetry and base cations by plasma emission spectroscopy. Trace elements (Cu, Zn, Fe, Mn) were determined by atomic absorption spectroscopy (AAS) after digestion with a mixture of perchloric and concentrated nitric acid. Boron was measured by colorimetry after dry ashing and digestion with hydrochloric acid. Quality control was ensured by the inclusion of standard reference materials in each batch of samples.

Minespoil chemistry

At each subsite, surface samples (0–150 mm) of minespoil weighing approximately 2·0 kg were taken from positions adjacent to each of the five assessment trees. Sample points were located at a distance of half the crown radius from the base of each stem, on the ridge crest in both upslope and downslope positions. These were bulked into two samples for each subsite. Minespoil pH, electrical conductivity, cation exchange capacity, loss-on-ignition, total nitrogen content and extractable concentrations of phosphorus, potassium, magnesium and manganese were performed on the < 2-mm fraction of air-dried minespoils at the Environment and Industry Research Unit, University of East London, UK.

Minespoil pH was measured in a 1 : 2·5 w/v suspension and electrical conductivity in an extract of calcium sulphate solution. Cation exchange capacity was determined using barium chloride as an electrolyte, and magnesium sulphate to exchange with barium and precipitate it out. Magnesium in solution was determined by titration with disodium EDTA (Bascomb 1964). Loss-on-ignition was calculated after samples had been placed in a furnace at 850 °C for 30 min (Ball 1964). Total nitrogen (Kjeldahl) concentrations were measured by colorimetry after digestion with concentrated sulphuric acid. Extractable phosphorus concentrations were determined by method two of Bray & Kurtz (1945). Samples were prepared using ammonium fluoride and hydrochloric acid, and measured by AAS. Extractable potassium and magnesium were measured by flame emission spectroscopy and AAS after extraction with M ammonium nitrate (Ministry of Agriculture, Fisheries & Food 1986).

Root mapping

Examination of tree root systems used the standard trench methods described by Böhm (1979) and Yeatman (1955). Root systems were excavated at the Abercrave 3, Bryn Pica 2, Dunraven 2 and Fyndaff 3 sites. At each, one or two representative trees were chosen in each subsite for detailed examination. Trenches were dug to 1 m depth using a pick and shovel. Each trench was located at right angles to plough ridges, 0·50 m from the base of the tree stem on its downslope side. Each root position was mapped using a grid, and the diameter measured using callipers. Maximum rooting depth, width, area and total number of roots exposed in the trench face were determined for each tree. The percentage of roots occupying each 100-mm layer within the profile, and cumulative totals with depth, were also calculated.

Soil physical measurements

An excavation–replacement technique for soil bulk density measurement was used (Smith & Thomasson 1974), which allowed consecutive volume measurements through each minespoil profile. In each trench, a hole approximately 0·3 × 0·3 m by 0·2 m deep was excavated and spoil removed for determination of weight, stone content and moisture content. The volume of the excavation was measured by backfilling to the original ground surface, using plastic spheres 20 mm in diameter. This sequence of operations was repeated for 0·2–0·4 m, 0·4–0·6 m and, at one site, 0·6–0·8 m depth intervals.

The sample weight of field moisture was determined on site using a spring balance. Stones > 20 mm in diameter were removed by wet sieving, which involved washing with a high pressure jet spray to remove adhering soil particles. A 2·0-kg subsample of the < 20 mm soil fraction was removed and moisture content determined. The total stone volume was finally subtracted from the original sample volume to give the volume occupied by soil fines (< 20 mm). Bulk density of the soil fines was calculated by dividing the weight (adjusted for stone and moisture contents) by the volume.

Soil hydrological monitoring

Because it was impossible to bore into minespoils with a high sandstone content, water table depth measurements were restricted to seven sites. At three positions within each subsite, boreholes 75 mm in diameter were excavated to a depth of 0·8 m. Each hole was lined with a plastic drainage pipe slotted at regular intervals. The tubes were capped to prevent direct ingress of rain. Boreholes were read weekly from 26 October 1988 to 17 October 1989. Results are expressed as the number of days mean water levels in the three boreholes occurred above 0·4 m and 0·7 m depth during the monitoring period (W40 and W70 days; Hodgson 1976). Wetness classes were assigned to each subsite according to the degree and duration of waterlogging conditions (Hodgson 1976).


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Growth of l. leptolepis

Height, age and estimates of yield class for L. leptolepis at each site are presented in Table 2. The largest trees were found in good subsites at the Abercrave 2 and 3 sites, where the yield class exceeded 11. Exceptionally slow growth was found in very poor subsites at the Abercrave 2 and Dunraven 2 sites, where yield class was less than 2. A yield class of about 10 is expected for Japanese larch on undisturbed ground in the locality (I. Blake, personal communication)

Table 2.  Age, height and estimates of yield class for Larix leptolepis at 10 minespoil sites in South Wales
SiteSubsiteAge (years)Mean height (m)Height range (m)Height SD (m)Yield class (m3 ha–1 year–1)
Abercrave 1Good197·616·68–8·950·945·6
Medium 5·704·59–6·450·643·6
Poor 4·423·31–5·130·672·2
Abercrave 2Good1710·9310·50–12·150·6211·2
Medium 8·477·53–9·670·707·8
Poor 5·835·06–6·460·504·6
Very poor 2·812·29–3·390·471·2
Abercrave 3Good1712·5811·70–13·250·5013·8
Medium 8·326·21–9·500·257·4
Poor 4·563·95–5·010·433·2
Bryn Pica 1Good147·216·15–7·720·578·0
Medium 6·555·00–7·841·097·0
Poor 5·424·34–6·420·815·6
Very poor 2·562·46–2·780·112·4
Bryn Pica 2Good146·565·95–6·950·357·0
Medium 4·552·69–6·261·184·4
Poor 2·281·96–2·560·211·8
Dunraven 1Good85·144·58–5·650·4311·0
Medium 3·032·04–3·510·546·2
Poor 2·031·41–2·640·494·2
Dunraven 2Good83·693·17–4·100·337·8
Medium 3·212·72–3·450·286·6
Poor 1·531·09–1·770·253·2
Very poor 0·670·35–1·090·281·2
Fyndaff 1 63·292·87–3·610·269·2
Fyndaff 2Good63·473·09–3·790·299·6
Medium 2·712·39–3·090·257·4
Poor 1·631·21–2·050·284·8
Fyndaff 3Good72·472·11–2·690·235·8
Medium 2·031·65–2·270·215·0
Poor 1·010·69–1·250·232·2

Foliar nutrient concentrations

A summary of foliar nutrient concentrations is given in Table 3. Using data presented in Stone (1968) and Van den Burg (1985), each observation was placed into one of five categories, representing high, optimum, normal, inadequate and deficient concentrations of each element. Concentrations of nitrogen were consistently below the normal range of published values, with over 80% of trees having concentrations within the deficient range. The lowest nitrogen content measured (0·50%) was lower than any reported by Van den Burg (1985). Phosphorus concentrations were within the normal range in over 80% of subsites, with only 7% within the deficient range. In contrast, concentrations of potassium and magnesium were often substantially greater than the normal range of published values. Concentrations of calcium were within the normal range in over 80% of subsites.

Table 3.  Nutrient concentration in Larix leptolepis sampled across the 10 study sites. For each nutrient, the first row classifies foliar concentrations according to adequacy. The second row shows the frequency distribution (%) of samples taken in this study
Nutrient concentration ranges
Nitrogen (%) > 3·003·00–2·602·59–1·801·79–1·60 < 1·60
Phosphorus (%) > 1·01·00–0·760·75–0·300·29–0·18 < 0·18
Potassium (%) > 1·81·80–1·421·41–0·750·74–0·50 < 0·5
Magnesium (%) > 0·240·24–0·200·19–0·140·13–0·10 < 0·10
Calcium (%) > 1·001·00–0·680·67–0·280·27–0·20 < 0·20
Iron (mg kg–1)– – 230–170 6·5169–50 93·5 < 50 0·0
Manganese (mg kg–1)25002500–15911590–372371–250 < 250
Zinc (mg kg–1)– – 76–44 16·243–15 83·8 < 15 0·0
Boron (mg kg–1)– – 69–28 38·727–15 61·3 < 15 0·0
Copper (mg kg–1)– 14·0–6·16·0–2·01·9–1·3 < 1·3

Concentrations of iron, zinc and boron were low compared with the normal range of published values. In contrast, manganese concentrations were exceptionally high, and more than 60% of trees had concentrations within the optimum and high ranges. Concentrations of copper were within the normal and optimum ranges in over 80% of subsites.

Correlation coefficients between yield class and foliar nutrients and some elemental ratios are shown in Table 4. Nitrogen (r = 0·73, n = 31, P < 0·001) and zinc (r = 0·70, n = 31, P < 0·001) were highly correlated with yield class, while significant but smaller correlations were obtained for phosphorus (r = 0·42, n = 31, P < 0·05) and the calcium/magnesium ratio (r = 0·44, n = 31, P < 0·05). Two statistically significant negative relationships occurred, with magnesium (r = –0·59, n = 31, P < 0·001) and boron (r = –0·36, n = 31, P < 0·01).

Table 4.  Pearson product–moment correlation matrix for foliar nutrients, element ratios, yield class (YC) and age of Larix leptolepis. Negative correlations are indicated below the leading diagonal and positive correlations above
  • *

    P < 0·05,

  • **

    P < 0·01,

  • ***

    P < 0·001.

YC1·00 0·73***0·42*0·08 0·13 0·290·70*** 0·34 0·050·44*
Age–0·191·00 0·48**0·060·270·64*** 0·250·12    0·27
N –0·081·000·20     0·62*** 0·31  0·24
P   1·000·45** 0·57*** 0·41*0·40*    0·48**
K  –0·14 1·00 0·26 0·110·25    0·21
Mg–0·59*** –0·28–0·17–0·231·00 0·31  0·61*** 0·32  
Ca  –0·04  –0·041·00 0·65***0·18     
Fe–0·16–0·31–0·11–0·31–0·11 –0·281·00  0·340·110·330·06 
Mn  –0·01  –0·32 –0·35*1·000·27 0·10   
Zn     –0·44** –0·11 1·00 0·31 0·110·67***
B–0·36*–0·12–0·15–0·43*–0·58*** –0·27 –0·29–0·42*1·000·030·59*** 0·33
Cu –0·09 –0·26–0·21–0·10–0·10    1·000·120·05 
N/P–0·09–0·34*  –0·53** –0·36* –0·33–0·17  1·00  
K/Ca –0·51**–0·02–0·13 –0·29 –0·27–0·42* –0·27 –0·171·00 
Ca/Mg          –0·34–0·02–0·18 1·00

Multiple regression analysis was used to establish a general model of response between yield class and nutrient elements (Table 5). Nitrogen, magnesium and phosphorus contributed to the model, which gave a highly significant regression (P < 0·001). Nitrogen and magnesium were the most significant variables. The adjusted R-squared value indicated that 73% of variability in the yield class was explained by the model.

Table 5.  Multiple regression model of yield class on foliar nutrients for Larix leptolepis
Independent variableEstimateStandard errort-valueSignificance level
  1. n = 31 observations fitted.


The relationship between yield class and foliar magnesium concentration was examined further using a distance-weighted least-squares procedure (Fig. 1). The figure demonstrated that yield class was severely depressed until a foliar concentration of < 0·25%, corresponding to the optimum concentration derived from nutritional studies (Table 3; Stone 1968; Van den Burg 1985).


Figure 1. Scatterplot of yield class against foliar magnesium concentration. Curve fitted according to distance-weighted least-squares.

Download figure to PowerPoint

Minespoil chemistry

The minespoil chemical properties are summarized in Table 6, and their relationship with each other and yield class are summarized in Table 7. Minespoil pH varied from strongly acid (pH 4·5) to moderately alkaline (pH 8·2), with a mean (pH 6·7) close to neutral. Electrical conductivity mirrored pH, with high values associated with alkaline samples. Cation exchange capacity was generally low. Loss-on-ignition values were higher than expected and this may have been caused by the combustion of coal and carbonaceous shale fragments included in the samples. Total minespoil nitrogen concentrations were comparable to those expected in native upland soils (Cooke 1967). Concentrations of extractable phosphorus were extremely variable, but generally low. Extractable potassium concentrations were moderate, while magnesium concentrations were high, at some sites exceptionally so (Ministry of Agriculture, Fisheries & Food 1986).

Table 6.  Minespoil analytical data. Samples taken from 10 study sites listed in Table 1
  • *

    Mean of 31 subsites;

  • †mean of 21 subsites;

  • ‡mean of 12 subsites.

Electrical conductivity (Ec) (mS cm–1)481941469935
Cation exchange capacity (CEC) (me 100 g–1)5·711·35·68·11·3
Loss-on-ignition (LOI) (%)1·75·33·62·91·1
Total nitrogen (% dry weight)0·060·290·230·16*0·05
Extractable phosphorus (mg kg–1)0·511·811·35·22·6
Extractable potassium (mg kg–1)8421212813640
Extractable magnesium (mg kg–1)384890505586128
Extractable manganese (mg kg–1)223702479491140
Table 7.  Pearson product–moment correlation matrix for soil chemical properties and yield class of Larix leptolepis. Negative correlations are indicated below the leading diagonal and positive correlations above
  • *

    P < 0·05,

  • **

    P < 0·01,

  • ***

    P < 0·001.

Yield1·00  0·34   0·57**0·18
pH–0·56**1·00  0·130·52*0·77***  
N–0·29–0·081·00  0·23 0·29 
P –0·32–0·011·000·100·01 0·310·14
K–0·29 –0·28 1·00   0·30
Mg–0·58**    1·00   
Ec–0·28 –0·05–0·57**–0·21 1·00  
CEC –0·77***  –0·45*–0·41–0·45*1·000·24
LOI –0·33–0·40  –0·25–0·37 1·00

Correlation analysis (Table 7) indicated significant positive relationships between pH and extractable magnesium or electrical conductivity. Cation exchange capacity was negatively correlated with extractable potassium, pH and electrical conductivity. Tree yield class was negatively correlated with pH and extractable magnesium, and positively correlated with cation exchange capacity.

Minespoil bulk density and stone content

A summary of minespoil bulk density and stone content measurements is shown in Table 8. Although all study sites had been cultivated to relieve compaction prior to tree planting, over 70% of all samples below 0·2 m exceeded 1·7 g cm–3. Stone contents often exceeded 25%, with a high proportion of stones greater than 100 mm in diameter. Bulk density and stone content increased markedly with depth across all subsites, especially between 0 and 0·2 m and 0·2–0·4 m depth. Significant positive relationships occurred between bulk density and depth (r = 0·41, n = 61, P < 0·001), and stone content and depth (r = 0·29, n = 61, P < 0·05), although a relationship between density and yield class could not be established.

Table 8.  Bulk density and stone content measurements from trenches excavated at the Abercrave 3, Bryn Pica 2, Dunraven 2 and Fyndaff 3 study sites
Bulk density (g cm–3)Stone content (%)
Depth of sampling increment (m)Number of samplesMeanRangeSDValues (%) greater than 1·7 g cm–3MeanRangeSDValues (%) greater than 25%
  • *

    Results for the 0·6–0·8 m increment refer only to the Abercrave 3 site.


Root distribution

Rooting depth was severely restricted for most trees that were investigated, with a maximum depth of 0·64 m and an average depth of 0·40 m (Table 9). Across all sites studied the mean proportion of roots within 0·1 m and 0·3 m of the minespoil surface were approximately 35% and 90% of the total number exposed. The total number of roots was determined by tree age, and there were correlations between root numbers and depth of rooting and tree height (r = 0·77 and 0·73, n = 19, P < 0·001, respectively). The correlation between tree age and rooting width (r = 0·75, n = 19, P < 0·001) was stronger than for rooting depth (r = 0·55, n = 19, P < 0·01), which suggests that lateral root extension was more important than vertical root penetration in root system development.

Table 9.  Rooting parameters of Larix leptolepis trees excavated at the Abercrave 3, Bryn Pica 2, Dunraven 2 and Fyndaff 3 study sites
SiteSubsiteAge (years)Tree height (m)Maximum rooting depth (m)Maximum rooting width (m)Rooting area (m2)Total number roots*
  • *

    Number of roots exposed in the trench face.

Abercrave 3Good (1)1713·060·641·980·92320
Good (2)1712·380·471·920·63282
Medium (1)178·470·481·870·61396
Medium (2)179·460·642·210·96440
Poor (1)174·780·491·220·45239
Poor (2)174·220·341·360·29162
Bryn Pica 2Good (1)147·920·381·280·2895
Good (2)146·940·421·650·50225
Medium (1)147·560·421·520·3693
Medium (2)145·750·431·310·30140
Poor (1)142·860·341·360·35101
Poor (2)142·510·321·280·28110
Dunraven 2Good83·850·611·150·3598
Very Poor81·040·201·110·1166
Fyndaff 3Good72·200·211·010·1845

Although it was not possible to establish a threshold at which the penetration of roots into minespoils was actually prevented, high bulk density and large stone contents were probably responsible for the restricted vertical development of the majority of root systems excavated. Much qualitative evidence of physical impediment to root penetration was found by observing root form. On descending, roots were often seen to branch profusely, or would turn and run horizontally. Roots were deflected around obstacles, with successive events producing severe deformation. Vertical roots sometimes thickened before tapering sharply, and occasionally ended in a mass of fine roots. The deepest roots followed channels or cracks. A significant correlation between yield class and rooting depth (r = 0·65, n = 19, P < 0·01) indicated that trees in stands of high productivity were more capable of developing deeper root systems than trees of lower yield class, or alternatively had been permitted to do so by more favourable site conditions.

Reconstructions of root distribution in cross-section were plotted for each tree excavated, and examples are presented for the Abercrave 3 good and poor subsites in Fig. 2. Marked differences in rooting habit occurred between trees in these two subsites, and these were reflected in all rooting parameters (Table 9). Almost all major roots (greater than 14 mm in diameter) in the good subsite emerged from the exposed face within 0·3 m of the minespoil surface. Major roots were often observed to turn outwards at a shallow depth from below the root collar at near right angles, and then run parallel to plough ridges. The strong preferred orientation of major roots parallel to the direction of cultivation indicated that the main structural root system was likely to be linear and asymmetrical.


Figure 2. Comparison of the distribution of Larix leptolepis roots from (a) good and (b) poor trees exposed in section at the Abercrave 3 site.

Download figure to PowerPoint

Borehole water table levels

Table 10 shows consistent differences between good and poor subsites at Abercrave and Bryn Pica. Poorer subsites were subject to more severe and prolonged periods of waterlogging than good subsites, the majority of which were rarely waterlogged at 0·7 m depth even during the winter period. However, the three sites at Fyndaff contrasted strongly, two rarely affected by water tables but Fyndaff 3 showing pronounced waterlogging throughout much of the monitoring period. For all the sites studies, a significant negative correlation occurred between maximum rooting depth and the number of days water tables were above 0·4 m (r = –0·55, n = 9, P < 0·05), with a significant positive correlation between maximum rooting depth and the number of days water table levels were below 0·8 m (r = 0·56, n = 9, P < 0·05).

Table 10.  Number of W40 and W70 days, and wetness classes at seven study sites (October 1988–October 1989)
SiteSubsiteTopographic positionSlope angle (°)W40 daysW70 daysWetness class*
  1. *Explanation of moisture wetness classes (Hodgson 1976):

  2. Wetness class I Not waterlogged to within 0·7 m depth for more than 30 days

  3. Wetness class II Waterlogged within 0·7 m for 30–90 days

  4. Wetness class III Waterlogged within 0·7 m for 90–180 days

  5. Wetness class IV Waterlogged within 0·7 m but not 0·4 m for more than 180 days

  6. Wetness class V The soil profile is wet within 0·4 m for more than 180 days and is usually wet within 0·7 m for more than 335 days

Abercrave 1GoodUpper slope32028I
MediumMiddle slope3270238IV
PoorLower slope3242175III
Abercrave 2GoodMiddle slope800I
MediumMiddle slope607I
PoorFlat (upper)4098III
Abercrave 3GoodMiddle slope400I
MediumLower slope200I
PoorFlat (upper)063308IV
Bryn Pica 2GoodMiddle slope18014I
MediumMiddle slope1800I
PoorMiddle slope180175III
Fyndaff 1GoodUpper slope15014I
Fyndaff 2GoodUpper slope3600I
MediumUpper slope3600I
PoorUpper slope3600I
Fyndaff 3GoodFlat (upper)2217336V
MediumMiddle slope127182IV
PoorFlat (lower)6210350V


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Nitrogen nutrition

Although total (Kjeldahl) nitrogen contents in minespoils were generally similar to native soils, foliar concentrations indicated widespread nitrogen deficiency. Sedimentary deposits, especially coal overburdens, can contain significant amounts of nitrogen (Reeder & Berg 1977a,b; Aldag & Strzyszcz 1980). However, the trees showed an inability to take up adequate amounts of spoil-derived nitrogen, suggesting that it is largely in unutilizable forms (Palmer, Morgan & Williams 1985). A shortage of plant-available nitrogen is unsurprising, given the very limited amount of soil material used in restoration of the opencast sites. The tree response, in the form of slow height growth, is typical of nitrogen deficiency on restored land (Bradshaw 1983; Kendle & Bradshaw 1992; Moffat & Buckley 1995). The study suggests that without inorganic or organic nitrogen amendment, reclamation of coal spoils to a larch forestry after use is unrealistic.

Magnesium nutrition

Very high concentrations of magnesium were recorded in the minespoils and in the L. leptolepis foliage sampled. The strong negative relationships between yield class and foliar magnesium concentration or minespoil extractable magnesium content suggest that growth was severely affected by excess of the element. Similar conclusions have been drawn elsewhere by Kulagin (1974) and Richardson & Evans (1987) for European larch L.. decidua Mill on magnesium-rich soils and minespoils, respectively.

The abnormally high concentrations of magnesium in the minespoils are probably derived from illite and mixed illite-smectite clay mineral groups in the shales and mudstones that comprise much of the overburden strata sequence (Davies & Bloxham 1974; Gill, Khalaf & Massoud 1977). Large amounts of plant-available magnesium are unusual in undisturbed or unpolluted British soils, but are common in soils developed over serpentine rocks, where they have been related to inhibition of plant growth (Whittaker 1954a,b; Proctor & Woodell 1971; Brooks 1987). Several mechanisms have been proposed to explain the effect of elevated magnesium on serpentine floras, including toxicity of magnesium itself and an unfavourable calcium/magnesium ratio (Walker 1954). Sulej et al. (1970) and Morgan, Jackson & Volk (1972) have also suggested that the assimilation of mineral nitrogen, and subsequently nitrogen metabolism, is reduced at low calcium/magnesium ratios.

The strong relationship between yield class and minespoil magnesium content suggests that purposefully selecting materials for restoration that contain the smallest amounts is desirable if larch remains a favoured species in minespoil reclamation. Bending (1993) found that shales contained much larger amounts of extractable magnesium than mudstone or sandstone, and this suggests that chemical characterization of the various lithologies available should take place before material selection.

The relationship between minespoil physical properties and rooting habit

Minespoils were characterized by their high bulk density and large volume stone contents, features that are also common to forested strip mine sites in North America and elsewhere (Pedersen, Rogowski & Pennock 1978, 1980; Bussler et al. 1984). Shallow tree rooting was found to be widespread, with the maximum rooting depth less than 0·65 m and the mean for all trees examined just exceeding 0·4 m. The mean proportion of roots within 0·1 m and 0·3 m of the minespoil surface was approximately 35% and 90% of the total number observed. There was a significant correlation (r = –0·46, n = 19, P < 0·05) between maximum rooting depth and minespoil bulk density at maximum rooting depth, in accordance with other studies (Torbert et al. 1988; Heilman 1990). In addition, qualitative evidence of physical impediment to root penetration was found by observing root form. These characteristics are consistent with those described by Bengough & Mullins (1990) in their study of the effects of mechanical impedance on root growth.

The pattern of disturbance created by ground cultivation appeared to affect the horizontal development of root systems profoundly. The majority developed in the horizontal plane, along the axis of the plough ridge, the amplitude of which appeared sometimes to provide an obstacle to lateral root extension. The development of asymmetrical root systems resulting from ground cultivation practices has also been observed on conventional forest sites and linked with wind instability (Yeatman 1955; Savill 1976; Deans 1983).

Waterlogging and its effects on growth

High water tables were present from late autumn to mid-spring at many of the subsites. Waterlogging was most frequent on flat areas, probably due to the lack of drainage gradient and minespoil compaction, which limited both lateral flow and vertical percolation. Inappropriate ground cultivation also caused surface ponding when ploughed ridges intersected one another or where furrows were especially deep. ‘Contour ploughing’ produced particularly unsatisfactory results, encouraging ponding behind ridge walls and restricting water movement downslope (Broad 1979). There was a rapid response of water table levels to major rainfall events, confirmed statistically for some sites by the stronger correlation of levels with rainfall in the 24-h period prior to monitoring, than to longer time periods. Weekly fluctuations in water table levels of up to 0·35 m were recorded at several subsites.

Vis (1974) found that growth of Japanese larch plantations in the Netherlands was best on sites where seasonal water table fluctuations ranged from below 0·4 m to above 0·8–1·2 m from the surface. In the present study, 11 subsites (58% of total) displayed water levels within the limits of this range (equivalent to wetness classes I and II). Tree performance at these subsites was medium or good (yield class mean = 6·6, SD = 3·2). Intermediate drainage conditions (equivalent to wetness classes III and IV) prevailed at six subsites (32%) and the majority of trees showed poor or medium growth (yield class mean = 3·2, SD = 1·2). The difference in yield class between subsites with drainage conditions of wetness classes I/II and III/IV was statistically significant (t = 2·5, n = 17, P < 0·05), strongly suggesting that waterlogging is extremely important in depressing growth rates locally on restored ground. The occurrence of shallow, near stationary, water tables during periods of active growth in the early spring is likely to have been detrimental to root function (Russell 1977).

Revised reclamation practice

The general infertility of minespoils revealed by our study emphasizes the need to save all suitable soils and drift deposits in the course of site working, and to allocate these appropriately during restoration. However, shortfalls are likely to occur on sites that boast little original soil cover, or have been subjected to previous injudicious coal extraction. There, opportunities may exist to select soil-forming materials from geological strata above or between coal-bearing layers (Daniels & Amos 1984). The study has shown that some minespoils contain large amounts of plant-available magnesium, and their identification and rejection for restoration purposes could increase the prospects for successful forestry reclamation.

The low organic matter and plant-available nitrogen contents in the minespoils investigated suggest that the use of organic amendments would be potentially useful. Various forms of sewage sludge and compost materials have been tested during the 1990s (Moffat, Bending & Roberts 1991; Bending & Moffat 1997a,b) and found to be effective in remedying nitrogen deficiencies and promoting tree growth. In the absence of proper soil materials for use in restoration, commercial forestry would seem to be unrealistic without such amelioration.

The study demonstrates that it is important to design new landforms with sufficient gradient to promote drainage, and to avoid compaction so that tree roots can penetrate spoils deeply. Despite spoils having been cultivated, compaction was commonplace, but recent research has shown that careful spoil and soil placement can prevent compaction from taking place (Bending 1993; Moffat & Boswell 1997). It is also vital to choose tree species that are suited to the prevailing edaphic and climatic conditions. Site survey, including soil analysis, is an essential step in planning new woodland on disturbed land.

Our study has shown that an understanding of the physical and chemical factors affecting tree performance on restored opencast coal spoils in South Wales has enabled beneficial changes to reclamation practice, and renewed the prospect of commercial forestry on restored land.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank British Coal, in particular Ian Carolan, for funding the research reported here, and John Roberts for fieldwork assistance. Ernest Ward was responsible for foliar analysis. Dr Richard Jinks made helpful comments on a previous version of this paper.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Aldag, R.W. & Strzyszcz, Z. (1980) Inorganic and organic nitrogen compounds in carbonaceous phyllosilicates on spoils with regard to forest reclamation. Reclamation Review, 3, 69 73.
  • Ball, D.F. (1964) Loss-on-ignition as an estimate of organic matter and organic carbon in non-calcareous soils. Journal of Soil Science, 15, 84 92.
  • Ballard, T.M. & Carter, R.E. (1985) Evaluating Forest Stand Nutrient Status. Land Management Report 20. British Columbia Ministry of Forests, Victoria, Canada.
  • Bascomb, C.L. (1964) Rapid method for the determination of cation-exchange capacity of calcareous and non-calcareous soils. Journal of the Science of Food and Agriculture, 15, 821 823.
  • Bending, N.A.D. (1993) Site factors affecting tree response on restored opencast ground in the South Wales Coalfield. PhD Thesis. University of East London, London, UK.
  • Bending, N.A.D. & Moffat, A.J. (1997a) The use of wood residues as amendments to coal minespoils for forestry reclamation in Wales, UK. Proceedings of the Forest Alternative for the Use of Residuals, Seattle, 1997.
  • Bending, N.A.D. & Moffat, A.J. (1997b) The use of thermally dried sludge as an organic amendment to coal minespoils: experience in the establishment of amenity and forest trees on reclamation sites in South Wales, UK. Proceedings of the 2nd European Biosolids and Organic Residuals Conference, Wakefield, 17–19 November 1997.
  • Bengough, A.G. & Mullins, C.E. (1990) Mechanical impedance to root growth: a review of experimental techniques and root growth responses. Journal of Soil Science, 41, 341 358.
  • Böhm, W. & Mullins, C.E. (1979) Methods of Studying Root Systems. Springer-Verlag, Heidelberg, Germany.
  • Bradshaw, A.D. (1983) The reconstruction of ecosystems. Journal of Applied Ecology, 20, 1 17.
  • Bray, R.H. & Kurtz, L.T. (1945) Determination of total, organic, and available forms of phosphorus in soils. Soil Science, 59, 29 45.
  • Broad, K.F. (1979) Tree Planting On Man-Made Sites in Wales. Forestry Commission Occasional Paper 3. Forestry Commission, Edinburgh, UK.
  • Brooks, R.R. (1987) Serpentine and its Vegetation: A Multidisciplinary Approach. Crook Helm, London, UK.
  • Bussler, B.H., Byrnes, W.R., Pope, P.E., Chaney, W.R. (1984) Properties of minesoil reclaimed for forest land use. Soil Science Society of America Journal, 48, 178 184.
  • Cooke, G.W. (1967) The Control of Soil Fertility. Crosby Lockwood, London, UK.
  • Daniels, W.L. & Amos, D.F. (1984) Generating productive topsoil substitutes from hard rock overburden in the Southern Appalachians. Environmental Geochemistry and Health, 7, 8 15.
  • Davies, M.M. & Bloxham, T.W. (1974) The geochemistry of some South Wales coals. Upper Palaeozoic and Post-Palaeozoic Rocks of Wales (ed. T.R.Owen), pp. 225 262. University of Wales Press, Cardiff, UK.
  • Deans, J.D. (1983) Distribution of thick roots in a Sitka spruce plantation 16 years after planting. Scottish Forestry, 37, 17 31.
  • Department of the Environment (1991a) Survey of Derelict Land in England, 1988. HMSO, London, UK.
  • Department of the Environment (1991b) Survey of Land for Mineral Workings in England, 1988. HMSO, London, UK.
  • Everard, J.E. (1974) Fertilizers in the Establishment of Conifers in Wales and Southern England. Forestry Commission Booklet 41. HMSO, London, UK.
  • Gill, W.D., Khalaf, F.I., Massoud, M.S. (1977) Clay minerals as an index of the degree of metamorphism of the carbonate and terrigenous rocks in the South Wales coalfield. Sedimentology, 24, 675 691.
  • Grimshaw, P.N. (1992) Sunshine Miners: Opencast Coalmining in Britain 1942–92. Spellman and Walker, Bradford, UK.
  • Hamilton, G.J. & Christie, J.M. (1973) Influence of Spacing on Crop Characterization and Yield. Forestry Commission Bulletin 52. HMSO, London, UK.
  • Heilman, P.E. (1990) Growth of Douglas fir and Red alder on coal spoils in western Washington. Soil Science Society of America Journal, 54, 522 527.
  • Hodgson, J.M. (1976) Soil Survey Field Handbook. Soil Survey Technical Monograph 5. Soil Survey of England and Wales, Harpenden, UK.
  • Kendle, A.D. & Bradshaw, A.D. (1992) The role of soil nitrogen in the growth of trees on derelict land. Arboricultural Journal, 16, 103 122.
  • Kulagin, Y.Z. (1974) Phytotoxicity of soils affected by industrial fumes in the Urals. Rast. I Prom. Sreda, 3, 56 64[in Russian].
  • Mayhead, G.J., Broad, K., Marsh, P. (1974) Tree growth in the South Wales coalfield. Forestry Commission Research and Development Paper 108. Forestry Commission, Edinburgh, UK.
  • Ministry of Agriculture, Fisheries and Food (1986) The Analysis of Agricultural Materials. MAFF Reference Book 427. HMSO, London, UK.
  • Moffat, A.J. & Boswell, R.C. (1997) The effectiveness of cultivation using the winged tine on restored sand and gravel workings. Soil and Tillage Research, 40, 111 124.
  • Moffat, A.J. & Buckley, G.P. (1995) Soils and restoration ecology. The Ecology of Woodland Creation (ed. R.Ferris-Kaan), pp. 75 90. J. Wiley & Sons, Chichester, UK.
  • Moffat, A.J. & McNeill, J.D. (1994) Restoring Disturbed Land for Forestry. Forestry Commission Bulletin 110. HMSO, London, UK.
  • Moffat, A.J., Bending, N.A.D., Roberts, C.J. (1991) The use of sewage sludge as a fertiliser in the afforestation of opencast coal spoils in South Wales. Land Reclamation. An End to Dereliction? (ed. M.C.R.Davies), pp. 347 356. Elsevier, London, UK.
  • Morgan, M.A., Jackson, W.A., Volk, R.J. (1972) Nitrate absorption and assimilation in ryegrass as influenced by calcium and magnesium. Plant Physiology, 50, 485 490.
  • Palmer, J.P., Morgan, A.L., Williams, P.J. (1985) Determination of the nitrogen composition of colliery spoil. Journal of Soil Science, 36, 209 217.
  • Pedersen, T.A., Rogowski, A.S., Pennock, R. (1978) Comparison of morphological and chemical characteristics of some soils and minesoils. Reclamation Review, 1, 143 156.
  • Pedersen, T.A., Rogowski, A.S., Pennock, R. (1980) Physical characteristics of some minesoils. Soil Science Society of America Journal, 44, 321 328.
  • Proctor, J. & Woodell, S.R.J. (1971) The plant ecology of serpentine. I. Serpentine vegetation of England and Scotland. Journal of Ecology, 59, 375 395.
  • Reeder, J.D. & Berg, W.A. (1977a) Plant uptake of indigenous and fertilizer nitrogen from a Cretaceous shale and coal mine spoils. Soil Science Society of America Journal, 41, 919 921.
  • Reeder, J.D. & Berg, W.A. (1977b) Nitrogen mineralization and nitrification in Cretaceous shales and coal mine spoils. Soil Science Society of America Journal, 41, 922 947.
  • Richards, Moorehead & Laing Ltd (1996) Review of Landscaping, Restoration and Revegetation of Colliery Spoil Waste Tips and Lagoons. The Stationery Office, London, UK.
  • Richardson, J.A. & Evans, M.E. (1987) Establishing trees and shrubs on magnesian limestone waste. Journal of Environmental Management, 24, 157 167.
  • Russell, R.S. (1977) Plant Root Systems: Their Functions and Interaction with the Soil. McGraw-Hill International Book Company, New York, NY.
  • Savill, P.S. (1976) The effects of drainage and ploughing of surface water gleys on rooting and windthrow of Sitka spruce in Northern Ireland. Forestry, 49, 133 141.
  • Smith, P. & Thomasson, A.J. (1974) Density and water-release characteristics. Soil Survey Laboratory Methods (eds B.W.Avery & C.L.Bascomb), pp. 42 56. Soil Survey Technical Monograph 6. Soil Survey of England and Wales, Harpenden, UK.
  • Stone, E.L. (1968) Microelement nutrition of forest trees: a review. Forest Fertilization: Theory and Practice, pp. 132 175. National Fertilizer Development Center, Tennessee Valley Authority, Muscle Shoals, AL.
  • Sulej, J., Slesak, E., Leonowicz-Babiak, K., Buczek, J. (1970) Tentative explanation of dwarfish growth of plants on serpentine soils. I. Physicochemical and biological properties, and mineral elements of serpentine soils. Acta Societatis Botanicorum Poloniae, 39, 405 419[in Polish].
  • Torbert, J.L., Tuladhar, A.R., Burger, J.A., Bell, J.C. (1988) Minesoil property effects on the height of ten-year-old White pine. Journal of Environmental Quality, 17, 189 192.
  • Van den Burg, J. (1985) Foliar analysis for determination of tree nutrient status – a compilation of literature data. Rijksinstituut voor Onderzoek, in de bos-en Landschapsbouw, Rapport 414. Wageningen, NL.
  • Vis, T. (1974) Investigation of the relations between forest and soil on humus podzol soils. Nederlands Bosbouw Tijdschrift, 46, 94 111.
  • Walker, R.B. (1954) Factors affecting plant growth on serpentine soils. Ecology, 35, 259 266.
  • White, J. (1959) Afforestation of a former opencast site in Coed Morgannwg. Journal of the Forestry Commission, 28, 69 74.
  • Whittaker, R.H. (1954a) The ecology of serpentine soils. Ecology, 35, 258 259.
  • Whittaker, R.H. (1954b) The vegetational response to serpentine soils. Ecology, 35, 275 288.
  • Yeatman, C.W. (1955) Tree Root Development on Upland Heaths. Forestry Commission Bulletin 21. HMSO, London, UK.