• Open Access

Mobilization of minerals and moisture loss during senescence of the energy crops Miscanthus×giganteus, Arundo donax and Phalaris arundinacea in Wales, UK


R. Smith, tel. +44 1597 860373, fax +44 1597 860381, e-mail: smithR3@cardiff.ac.uk


Miscanthus×giganteus, Arundo donax and Phalaris arundinacea were planted at three sites in Wales, UK, to study the timing of mineral and moisture decline during autumn and winter. The mineral and moisture content at several harvest dates throughout the period were compared with power station threshold levels. All crops showed a highly significant decrease in moisture content in all years with the exception of Arundo which did not lose moisture in the final year. The Miscanthus and Arundo crops remained above the moisture content threshold in all years and would require further drying before combustion. The Phalaris fell below the moisture threshold in January. The Miscanthus leaves fell below the nitrogen threshold before leaf abscission at one site but not at another, although the cane remained mostly within the guideline range. Both its leaves and cane remained above the potassium threshold level. Its leaf sulphur fell within acceptable levels during November and its cane remained below the lowest threshold from November onwards. The Arundo leaves remained above the nitrogen, potassium and sulphur thresholds throughout the period and it did not lose its leaves. The cane nitrogen increased above the threshold during autumn whereas its potassium remained higher than its threshold, with the exception of one sampling date at one site. Its sulphur remained within the guidelines. The Phalaris crop fell to within nitrogen guidelines at one site but increased at the other. Its potassium decreased to below the threshold at both sites in January, and its sulphur content was consistently within the guidelines.


It is generally understood that the perennial energy crops Miscanthus×giganteus, Arundo donax and Phalaris arundinacea translocate nutrients from aboveground plant matter to the rhizomes during senescence. This occurs during late autumn and winter in Wales, UK. These nutrients are then remobilized during shoot elongation the following year, as suggested by Himken et al. (1997), who showed that phosphorus, potassium and magnesium concentrations in Miscanthus rhizomes decreased during May and June to coincide with shoot elongation. This reduces the requirement for fertilizer addition, thus reducing both financial and environmental costs.

In addition, as a result of translocation, a delayed harvest generates combustible material containing minimal nutrients. This improves the combustion quality of the crop by reducing corrosion, slagging, fouling and environmentally harmful emissions (Miles et al., 1996). High concentrations of potassium within the plant material decreases the ash melting point and contributes to corrosion and slagging (Obernberger, 1998), thus reducing the quality of the crop as a fuel. In contrast, calcium and magnesium can contribute to increasing the ash melting point (Obernberger, 1998), and therefore a high magnesium and calcium concentration combined with a low potassium content is most suitable. Nitrogen content is important when considering crop quality, as NOx emissions need to be minimized during combustion. Similarly sulphur affects the combustion process and can cause corrosion, as well as producing the pollutants SO2 and SO3 (Obernberger, 1998).

Some studies have estimated the threshold levels of these minerals in harvested material for use as a combustion crop. The guiding levels referred to in this study are 0.1% (Lewandowski & Kicherer, 1997) to 0.6% (Obernberger, 1998) nitrogen; 0.2% potassium (Jorgensen, 1997); and 0.1% (Obernberger, 1998) to 0.3% (Lewandowski & Kicherer, 1997) sulphur, of total dry weight. The guidelines for magnesium and calcium suggest that the ash should contain over 2.5% and 15–35%, respectively (Obernberger, 1998) in order to increase its melting point and prevent some combustion problems. There are no known combustion problems associated with phosphorus content. However, this mineral has been considered in this study because it is an essential nutrient that needs to be recycled in order to maintain low fertilizer requirements.

A further benefit of a delayed harvest is reduced moisture content, caused by the standing plant matter drying out naturally over the winter period. This reduces the need for artificial crop drying before storage or combustion. Moisture content of <23% is necessary to mitigate the danger of self-ignition during straw storage (Clausen, 1994), although El Bassam & Huisman (2001) suggests that this occurs in Miscanthus crops at moisture contents over 50%. Straw bales with high moisture content can also cause inefficiency within the power station as they are too heavy for the conveyor belt (J. Ariss, Elean Power Station, personal communication). Elean Power Station in Cambridgeshire, UK, which is the world's largest straw burning power station, rejects all bales with moisture content higher than 25%.

The aims of this study were to determine the timing of mineral and moisture decline relevant to using the crops Miscanthus, Arundo and Phalaris as combustible material, when grown in Wales, UK. It investigates the nitrogen, potassium, phosphorus, magnesium, calcium and sulphur concentrations during autumn and winter, and relates these to threshold levels for combustion, where known. It also calculates the crop moisture content during this period and relates this to moisture threshold levels.


Site details

The research was carried out at three sites within Wales, UK; Llwynprenteg, Bluestone and Llysdinam. Site 1, Llwynprenteg, is situated near Pont-Rhyd-y-Groes, Ceredigion (SN 687716), at an altitude of 90 m above sea level (a.s.l.). The soil type is typical brown earth of Denbigh 1 association, which is a brown, stony, well-drained soil. Herbicide was applied as glyphosphate at 5 L ha−1, then the field was ploughed, limed at 4 t ha−1, rolled and harrowed before planting sections with Miscanthus×giganteus rhizomes at 40 000 plants ha−1 and P. arundinacea seed at 7.5 kg ha−1, during March and April 2004. Fertilizers were then applied as inorganic N : P : K at 60 : 50 : 50 kg ha−1. The Phalaris section was reseeded during September 2004 following poor initial growth. The A. donax section was planted as rhizomes at 40 000 plants ha−1 during April 2005.

Site 2, Bluestone, consists of two fields near Narberth, Pembrokeshire (SN 063129 and SN 032149) at altitudes of 75 and 60 m a.s.l.. Both sites consisted of typical brown earths, the first field of the Milford association, which is a reddish, fine, loamy soil, and the second of the Denbigh 1 association, as above. Both sites were prepared with herbicide as described above, ploughed and rotivated, 1 month before planting. The first field was planted with Miscanthus rhizomes at 40 000 plants ha−1 during April 2003. The second field was planted with Arundo rhizomes at 40 000 plants ha−1 during April 2005.

Site 3, Llysdinam, is at Newbridge-on-Wye, Powys (SO 005581) at 190 m asl. The soil type is cambic stagnogley soil of the Cegin association, which is loamy, clayey and subject to seasonal waterlogging. The site was prepared during April 2004 with herbicide as described above, ploughed, rotivated and harrowed. The Miscanthus section was planted at 40 000 rhizomes ha−1 during April 2004 and the Phalaris section was planted at 7.5 kg ha−1 during May 2004 and reseeded during September 2004 following poor initial growth.

Site visits

During 2005–2006, the Miscanthus and Arundo crops at Llwynprenteg were sampled monthly from November to February.

During 2006–2007, all the crops at Llysdinam, Llwynprenteg and Bluestone were visited fortnightly from October to February. Miscanthus and Arundo crops were subsampled every 2 weeks and Phalaris was subsampled every 4 weeks. Phalaris sampling at Llwynprenteg ceased in January 2007, as the landowners were keen to harvest the crop. At Bluestone access problems caused disruption to Miscanthus subsampling during January and cessation of the Miscanthus study on February 2, 2007.

During the winter of 2007–2008, the Miscanthus crop at Llysdinam and Llwynprenteg, and the Arundo crops at Llwynprenteg and Bluestone were subsampled once during the first week of December, and again during the last week of February. The Miscanthus crop at Bluestone was dug up during 2007 and therefore sampling was not possible.

Crop subsampling

Miscanthus and Arundo moisture content. Five plants were randomly selected from the crop, for which all shoots were chopped at the base and collected. Five shoots were randomly selected from each plant, weighed fresh, oven-dried at 80 °C to constant weight, and reweighed dry. Crop moisture content was calculated.

Phalaris moisture content. All shoots within five randomly selected 1 m2 quadrats were cut at the base and collected. Fresh weight per 1 m2 area was determined immediately and a subsample of approximately 500 g from each area was accurately weighed fresh, oven-dried at 80 °C to constant weight, and reweighed dry. Crop moisture content was calculated.

Calorific value determination. Three oven-dried shoots sampled from each of the Miscanthus, Arundo and Phalaris crops at Llysdinam, Llwynprenteg and Bluestone during late November 2006 and late January 2007, were milled to powder and sent to the IBERS laboratories, Wales, for determination of calorific value.

Mineral content determination. During 2005–2006, five oven-dried shoots from each sampling date of the Miscanthus and Arundo crops at Llwynprenteg were sent to Direct Laboratories, Wolverhampton, for analysis of total nitrogen, phosphorus, potassium, calcium, magnesium and sulphur content. During 2006–2007, three leaf samples and three cane samples from each sampling date of the Miscanthus and Arundo crops at Llwynprenteg and Bluestone were sent to Natural Resource Management Ltd., Berkshire, for laboratory analysis of the same minerals. Three samples of whole aboveground plant matter from the Phalaris crops at Llwynprenteg and Llysdinam, from each sampling date, were also sent for analysis.

Statistical analysis of results

Analysis of crop moisture content and calorific value data. The moisture content and calorific value from the earliest sampling date were tested for differences from the data from the last sampling date using t-tests, for each crop and site. Data were transformed as necessary to meet the assumptions of the test and any which did not concur after transformation were tested using the nonparametric Mann–Whitney test.

Analysis of plant mineral content data. All datasets were analysed for differences between dates using analysis of variance. Data were transformed as necessary to meet the assumptions of the test and any which did not concur after transformation were tested using the nonparametric Kruskal–Wallis test.


Moisture content

In 2005/2006, the moisture content remained above 25% for all crops throughout the period (Fig. 1a). The lowest moisture value occurred during January for all crops except the Miscanthus at Llysdinam, which continued to decrease steadily from January to April. All other crops showed a slight increase in moisture at their final sampling date in February, although there was still a highly significant decrease for all sites between November and February (P<0.000).

Figure 1.

 Crop moisture content of Miscanthus (MG) at Llysdinam (LS), Llwynprenteg (LP) and Bluestone (BS); Arundo (AD) at Llwynprenteg (LP) and Bluestone (BS); and Phalaris (PA) at Llysdinam (LS) and Llwynprenteg (LP); in (a) 2005/2006, (b) 2006/2007 and (c) 2007/2008. Threshold value of 25% is shown.

In 2006/2007, the moisture content remained above 25% for all crops, with the exception of the late January sampling date for the Phalaris at both sites, although the Phalaris at Llysdinam regained moisture thereafter (Fig. 1b). The lowest moisture value occurred during late January or February for all crops. The difference between the earliest sampling date and the latest sampling date was highly significant for all Miscanthus crops (P<0.000), as was the difference between the November and January sampling dates for both Phalaris crops (P<0.000). The Arundo crops also showed a highly significant difference between first and last sampling dates at Llwynprenteg (P<0.000), and at Bluestone (P=0.006).

In 2007/2008, all Miscanthus crops contained less moisture in February than in December, which was highly significant at Llysdinam (P=0.001) and at Llwynprenteg (P<0.000). Despite this, all remained above the 25% threshold (Fig. 1c). The Arundo crops; however, did not lose any moisture in 2007/2008.

Mineral content

Figure 2 shows the mean mineral content for each crop and Table 1 lists the results of the statistical analysis for each dataset.

Figure 2.

 Mineral content of plant dry matter: (a) nitrogen, (b) phosphorus, (c) potassium, (d) calcium (e) magnesium (f) sulphur, of Miscanthus (MG) and Arundo (AD) whole plant matter in 2005/2006; Miscanthus and Arundo cane (ca) and leaves (lvs) in 2006/2007 at Bluestone (BS) and Llwynprenteg (LP); Phalaris (PA) whole plant matter in 2006/2007 at Llwynprenteg (LP) and Llysdinam (LS). Threshold levels are shown for (c) potassium, and upper and lower threshold levels are shown for (a) nitrogen and (f) sulphur.

Table 1. anova results for all mineral data
  anova results    anova results    anova results  
  1. Sampling dates in 2005–2006 (0506) for Miscanthus and Arundo crops at site Llwynprenteg (LP) only. Sampling from 2006 to 2007 (0607) for cane and leaves (Lvs) of Miscanthus, Arundo and whole plant data for Phalaris at Llwynprenteg (LP), Miscanthus and Arundo at Bluestone (BS), and Phalaris at Llysdinam (Llys).

  2. Date of highest mean value (HMV) and lowest mean value (LMV) are provided.

  3. ns, not significant; na, not available.

0506N 0.00025.22912/052/060506N 0.00010.32912/052/06na      
LP       LP              
 P 0.0016.52912/052/06 P 0.0044.7291/062/06       
 K 0.0016.22911/051/06 K 0.0054.5291/062/06       
 Ca 0.00067.82912/052/06 Ca 0.0007.62911/052/06       
 Mg 0.0016.12912/052/06 Mg 0.0015.92911/052/06       
 S 0.00013.42911/052/06 S 0.00012.22911/052/06       
LP Cane0.0392.72612/0610/06LP Cane0.0024.03511/0610/06LP      
 PLvs0.000165310/0612/06 PLvs0.0013.39310/062/07 P0.0027.6179/061/07
  Cane0.0016.52510/061/07  Cane0.0024.03512/069/06       
 KLvs0.001na510/0612/06 KLvs0.0008.39310/061/07 K0.00032.1179/061/07
  Cane0.000102510/061/07  Cane0.4 ns  10/061/07       
 CaLvs0.7 ns  12/069/06 CaLvs0.0004.09212/062/07 Ca0.0027.61711/061/07
  Cane0.1 ns  12/061/07  Cane0.1 ns  12/0612/06       
 MgLvs0.0133.3539/0612/06 MgLvs0.1 ns  10/062/07 Mg0.0423.31711/061/07
  Cane0.1 ns  10/061/07  Cane0.1 ns  9/0612/06       
 SLvs0.6 ns  12/0611/06 SLvs0.0005.09312/062/07 S0.0293.7179/061/07
  Cane0.2 ns  12/061/07  Cane0.0006.13511/061/07       
0607NLvs0.5 ns  11/0611/060607NLvs0.7 ns  10/062/070607N0.00019.01710/062/07
BS Cane0.0173.6231/0712/06BS Cane0.0024.33211/0610/06Llys      
 PLvs0.3 ns  11/0611/06 PLvs0.3 ns  10/062/07 P0.00074.41710/062/07
  Cane0.0016.22310/061/07  Cane0.0432.43211/0610/06       
 KLvs0.1 Ns  10/0611/06 KLvs0.0162.57211/061/07 K0.00092.11710/062/07
  Cane0.6 ns  10/061/07  Cane0.1 ns  11/061/07       
 CaLvs0.4 Ns  11/0611/06 CaLvs0.0005.97210/061/07 Ca0.00019.91710/062/07
  Cane0.0009.12310/062/07  Cane0.2 ns           
 MgLvs0.3 Ns  11/0611/06 MgLvs0.0005.17210/062/07 Mg0.00039.91710/062/07
  Cane0.0008.62310/0612/06  Cane0.3 ns  2/071/07       
 SLvs0.00010.13310/0611/06 SLvs0.0005.97212/061/07 S0.000221.51710/061/07
  Cane0.00038.12310/0612/06  Cane0.0182.93210/062/07       

Nitrogen (Fig. 2a)

The nitrogen contents of Miscanthus and Arundo leaves were higher than their canes, although the Miscanthus leaves decreased rapidly during November before leaf abscission. The nitrogen content of the Arundo leaves also showed a decline over the sampling period, although this was only significant at Llwynprenteg, and the leaves remained intact at both sites. Both the Arundo and Miscanthus cane showed fluctuations in nitrogen content, although the Arundo cane contained greater nitrogen than the Miscanthus cane overall. The latter remained within the threshold range and contained the lowest values of all crops. The nitrogen content of Phalaris was slightly higher than the threshold during October to December, thereafter the crop at Llwynprenteg showed a gradual but significant increase and the crop at Llysdinam showed a highly significant decrease to within the threshold range.

Phosphorus (Fig. 2b)

The Miscanthus leaves contained the most phosphorus of all the crops but the level decreased during November before leaf abscission, which was only significant at Llwynprenteg. The Miscanthus cane contained the least phosphorus of all crops. The phosphorus levels in the Arundo leaves declined during the period, which was significant only at Llwynprenteg. The phosphorus content of the cane fluctuated and did not show a particular decline throughout the period. The Phalaris at Llysdinam showed a highly significant dramatic decrease during the entire period whereas the crops at Llwynprenteg decreased significantly but less dramatically.

Potassium (Fig. 2c)

The highest potassium levels were shown for the Arundo crop at Llwynprenteg in both 2005/2006 and 2006/2007. The Arundo crops at Bluestone contained slightly lower potassium levels than the Miscanthus crops. It had consistently low levels in the cane which dropped below the threshold in January but then rose again in February, although the differences were not significant. The Miscanthus crops showed a significant decline in both leaves and cane at Llwynprenteg and a nonsignificant decline at Bluestone. Both Phalaris crops showed a highly significant steady decline, fell below the threshold early in January and remained low thereafter.

Calcium (Fig. 2d)

The leaves of both Miscanthus and Arundo were high in calcium. Miscanthus leaf abscission occurred during November/December which coincided with a highly significant decrease in calcium content of the Arundo leaves. The Arundo cane contained the lowest calcium levels which remained consistently low throughout the period. The calcium content of the Miscanthus cane decreased during the period, which was highly significant at Bluestone but not significant at Llwynprenteg. The Phalaris levels were also fairly low and showed a highly significant decline at both sites over the period. All crops contained a similar level of calcium by the end of February.

Magnesium (Fig. 2e)

The magnesium levels of the Arundo leaves fell during the period, which was highly significant only at Bluestone. They remained higher than all other crops throughout the period. The Arundo cane remained at a consistent level throughout the period and contained higher amounts than the Miscanthus cane and Phalaris crops. The Miscanthus leaves showed a significant decrease in magnesium before leaf abscission at Llwynprenteg but not at Bluestone, whereas the cane magnesium decreased highly significantly only at Bluestone.

Sulphur (Fig. 2f)

The sulphur content of the Arundo leaves was consistently higher than all other crops although it fell below the higher threshold in January at Bluestone and in February at Llwynprenteg, a decrease that was highly significant at both sites. The Arundo cane remained within the guideline range from November at Bluestone but remained constantly below this level at Llwynprenteg in 2005/2006 and 2006/2007, while showing a significant decline at both sites throughout the period. The Miscanthus leaves fell to within the threshold range at Bluestone before leaf abscission, whereas at Llwynprenteg it remained low prior to a sudden increase in December. The Miscanthus cane at Llwynprenteg remained below the lowest threshold level throughout the period and showed no significant fluctuations. At Bluestone, the cane showed a highly significant decline during October and thereafter remained below the lowest threshold value. The Phalaris at Llwynprenteg remained within the guideline range throughout the period whereas at Llysdinam the level fell significantly to below the lowest threshold in December.

Calorific values of all crops

No differences in calorific values between November and January sampling dates were found at Llysdinam or Bluestone, although the data from Llwynprenteg were significantly greater in January than November (P=0.039). The increase at Llwynprenteg was minimal (from 17.74 to 18.27 MJ kg−1). Similarly, the Arundo crops at Llysdinam, Bluestone and Llwynprenteg were analysed for differences in mean calorific value between late November sampling and late January sample dates, and no significant differences were found. The Phalaris crops at Llysdinam and Llwynprenteg were also analysed for differences between calorific values in late November and late January, and no significant differences were found.

The overall mean calorific values for each crop were: Miscanthus, 17.99 MJ kg−1; Arundo, 17.59 MJ kg−1; Phalaris, 17.94 MJ kg−1.


Moisture losses

The Miscanthus crops in 2005–2006 lost between 30% and 33% of their moisture between the November and February sampling dates, although the crop at Bluestone had lost 51% by the beginning of January and regained moisture during the month. These results are similar to reports from Hertfordshire, a low rainfall area in eastern England, where reductions were approximately 34% between October and February (Clifton-Brown et al., 2001).

The Miscanthus crops in 2006–2007 lost between 19% and 61% of their moisture between the November and February sampling dates. The lowest moisture loss (19%) was recorded at Llysdinam, and was also recorded by a trial in Ireland between the date of the first frost (October) and a late March harvest (Clifton-Brown & Jones, 2001). The values at Llysdinam decreased from November to January, and then increased in February, which was likely an effect of very high precipitation during January 2007, which may have been absorbed by the senesced plant material. A loss of 42% was recorded at Llwynprenteg, which is similar to losses recorded in Germany, of a 52% reduction between December and February (Lewandowski & Heinz, 2003). The Miscanthus crop at Bluestone lost 23% of its moisture between the end of November and the beginning of December, when access problems prevented further sampling. It is highly likely that the moisture would have continued to decrease through December. In the 2007–2008 sampling period, the moisture losses at Llysdinam (24%) and Llwynprenteg (49%) were similar to the previous year.

The Arundo crops showed a significant reduction in moisture content in 2005–2006; 23% and 14% at Bluestone and Llwynprenteg, respectively; and 2006–2007 (27% and 30%). However, in 2007–2008 no moisture was lost which suggests the growing season may extend in more mature crops and prevent senescence and drying out.

The Phalaris at both sites experienced the largest moisture losses of all crops in this study. The moisture content differences from November 2006 to January 2007 were 66% and 76% at Llwynprenteg and Llysdinam, respectively. These values are similar to those reported from Sweden, which were 75% moisture loss at northern sites and 71% at southern sites (Landstrom et al., 1996), and were greater than the 52% losses shown in Hertfordshire, UK (Christian et al., 2006).

Phalaris was the only crop that consistently dried out sufficiently to meet power station moisture content standards. The Miscanthus and Arundo crops would both require drying before combustion, which would potentially create additional economic and energy costs, although Lewandowski et al. (2003) showed field drying of Miscanthus to reduce both moisture and nutrient content. R. Luxton (personal communication) showed that harvested Arundo crops left to dry outdoors steadily lost moisture between March and August, by which time moisture content was below 25% (unpublished results), so this may be an inexpensive drying option for the crop.

Calorific value differences

The calorific value of all crops was unaffected by harvest date, with the exception of an increased value for the Miscanthus crop at Llwynprenteg for the second sampling date. This increase was minimal and was likely to be within the range of sampling variation.

Plant NPK content

The leaf nitrogen, phosphorus and potassium contents of Miscanthus fell rapidly before leaf abscission at Llwynprenteg, but not at Bluestone. The cane nitrogen and phosphorus fell gradually over the winter period at both sites, whereas the cane potassium only showed a significant decline at Llwynprenteg.

The majority of studies in the literature report a decline in aboveground Miscanthus nutrient content over winter (Himken et al., 1997; Jorgensen, 1997; Clifton-Brown & Jones, 2001; Clifton-Brown et al., 2001; Lewandowski & Heinz, 2003). However, studies that included early sample dates have found that nutrient levels peak in spring and early summer, and decline thereafter (Beale & Long, 1997), with the most rapid decline occurring between during the summer months (Jorgensen, 1997). Hertfordshire studies have suggested that flowering indicates the timing of nutrient movement within the plant, although Miscanthus rarely flowers in that area (Yates & Riche, 2007). Their study showed little variation in Miscanthus nutrient composition between November and February, and suggests that translocation may have occurred earlier. This study showed that, at these Welsh sites, many plants produced flowers throughout the winter period (unpublished results). The timing of flower emergence was not studied, but generally it occurred during September (personal observations). This would suggest that the plants had little time to translocate nutrients before the sampling period of this study; however, a detailed study of the plants' nutrient composition through the year would be necessary to clarify the process. In this study, sampling of the crop was carried out during the period October–February to coincide with visual signs of senescence, when the plants began to lose their green colouring.

Other studies have shown the decline in nitrogen within aboveground plant matter to be most likely due to translocation of the nutrients to the roots and rhizomes (Beale & Long, 1997) as it mostly occurs as a fixed element that is not leached during precipitation (Jorgensen, 1997). During this study, a slight increase in rhizome nitrogen was detected during the sampling period at Llwynprenteg, but no significant increase was shown at Bluestone (unpublished results), which suggests that the nitrogen may have been lost in leaf litter at the latter site. Potassium mostly occurs in plants as soluble ions, which are easily leached by precipitation (Jorgensen, 1997), and this is likely to be the primary cause of winter loss (Chapin et al., 1980). Bluestone received less rainfall than Llwynprenteg's nearest weather station during the winter of 2006–2007; therefore differing levels of precipitation may explain loss of potassium from plant matter at Llwynprenteg but not at Bluestone. Some translocation of potassium has been shown in other studies, as has translocation of phosphorus (Beale & Long, 1997), although studies of poplar suggest that phosphorus is mostly lost in leaf litter (Kauter et al., 2003). This study observed an increase in rhizome potassium content before leaf abscission at both Bluestone and Llwynprenteg (unpublished results), which suggests that this mineral was translocated from the aboveground plant matter. However, the increase in rhizome phosphorus at Llwynprenteg occurred subsequent to leaf abscission, which suggests that, at this site, it was possibly taken up from the leaf litter.

The samples collected from the Arundo crops showed no differences in leaf content of nitrogen, phosphorus or potassium throughout the winter period. The cane samples showed nitrogen content to increase at both sites, and phosphorus to increase at Llwynprenteg only. These results suggest that the crops continued to take up nutrients through the winter, and were not beginning to senesce. It has been shown in the Mediterranean that a proportion of this plant remains green during the winter period (El Bassam, 1998); this apparent lack of senescence in Arundo crops would cause problems for its use as an energy crop. The Phalaris crops at both Llwynprenteg and Llysdinam displayed a loss of potassium content in aboveground plant matter throughout the period from October until February. This was probably a result of leaching from the senescing crop, as discussed above, and was also found in other studies of Phalaris (Burvall, 1997; Christian et al., 2006). The phosphorus content of the Llwynprenteg crop decreased, but its nitrogen content increased over the period. At Llysdinam, where the harshest climate was experienced, the phosphorus and nitrogen content decreased during the winter. This suggests the two crops may have been at different stages of senescence, and that at Llwynprenteg the Phalaris was still taking up nitrogen for further plant growth. It is possible therefore that the Llwynprenteg crop continued to grow during the winter season which resulted in new shoots being included in the harvested material, as shown by other studies (Christian et al., 1999, 2006). This re-growth was sufficient at Llwynprenteg to increase mineral content levels above the threshold. The differences between the two sites were likely to be caused by decreased winter temperatures at Llysdinam, which resulted in less new shoot growth. Therefore this crop may be better suited to colder temperatures for energy crop purposes.

The majority of other studies have shown a decrease in Phalaris nutrient content during winter (Burvall, 1997; Hadders & Olsson, 1997; Yates & Christian, 2001; Christian et al., 2006), although Landstrom et al. (1996) showed the nitrogen decline to occur from August to October but not thereafter, and Yates & Riche (2007) recorded an initial nitrogen decline from July to October, followed by a lesser decline from November to January. The latter study suggested remobilization of nutrients probably occurred during late summer.

Combustion quality of plant material

The Miscanthus leaves contained more nitrogen than the threshold value before December, and both the cane and leaves had potassium values higher than the threshold. At Llwynprenteg, but not at Bluestone, the leaf nitrogen fell below the threshold before leaf fall. This supports the current practice to delay harvest of Miscanthus crops until after leaf abscission to avoid excessive mineral content in the harvested material. The potassium content remained higher than the threshold level throughout the period. In terms of sulphur, Miscanthus contained the lowest levels in its cane, with its leaves rapidly falling to acceptable levels during November. However, field drying of harvested Miscanthus has been shown to reduce its nutrient content (Lewandowski et al., 2003).

The Arundo leaves had higher nitrogen contents than the threshold level on all sampling dates. Monti et al. (2008) also showed Arundo to have increased leaf and cane nitrogen in comparison to Miscanthus. The fact that Arundo crops did not lose their leaves during the winter period may cause problems for its use as a fuel crop, as their high mineral content is unsuitable for combustion purposes. The nitrogen content of the Arundo cane in this study was initially below the highest threshold value, but increased thereafter and remained higher than the guideline range throughout the period. The leaves and canes also contained a higher potassium level than the threshold value on all dates, with the exception of February sampling at Bluestone. Arundo contained high amounts of sulphur in its leaves although the whole plant samples taken in 2005/2006 remained within the threshold guidelines. Further research is necessary to determine if the mineral levels decrease in later spring, in addition to study of the date of new shoot emergence, to determine if there is a time period more suitable for Arundo harvest. However, field drying of Arundo crops may also reduce its nutrient content, as has been shown for its moisture content (Luxton, unpublished results).

The Phalaris crops contained more potassium than the threshold level at both sites in October, and then fell below this level in January. Therefore, in terms of potassium, Phalaris was found to be the only crop suitable for combustion. At Llysdinam, the nitrogen content fell likewise, and was within the guideline range from January onwards. However, at Llwynprenteg, nitrogen levels increased throughout the winter period, and remained higher than the top threshold level. This crop clearly needs to be studied further to investigate its mineral cycling more fully. Other studies have shown Phalaris to have decreased potassium content at a January harvest in comparison to Miscanthus, but similar levels of nitrogen and phosphorus (Christian et al., 1999). Phalaris remained within or below the threshold guidelines in terms of sulphur content.

Increased calcium and magnesium can contribute to an increased ash melting point and therefore some content in the harvested material is beneficial. This study showed that the three crops under study do not benefit from high calcium, as the level was only high in Miscanthus and Arundo leaves before December. The three crops similarly did not contain high magnesium levels with the exception of the Arundo leaves before December.


The moisture content of Miscanthus and Arundo fell during the winter periods studied, although the amount lost varied considerably. The greatest moisture loss was recorded in the Phalaris crops and these were subsequently the only crops that achieved a final moisture content of less than 25%, thus being suitable for immediate combustion. Moisture losses from Miscanthus and Arundo were not sufficient to produce a crop suitable for immediate combustion, and would therefore require storage and further drying before combustion, when grown in the Welsh climate.

The most important factor influencing the timing of harvest of energy crops will be the mineral content thresholds as determined by power stations. The thresholds referred to in this study suggest that Miscanthus crops should be harvested from December onwards, after leaf abscission has occurred, although the cane still contained high levels of potassium. Arundo crops in this study did not meet nitrogen or potassium thresholds, and further investigation is required to determine if nutrient losses occur at a later date or during storage. The Phalaris at Llysdinam met the threshold levels from December onwards, but further study is required to determine the reason for the nitrogen increase recorded at Llwynprenteg. However, this study relates to Miscanthus plants aged 3–5 years old, Arundo plants 2–3 years old and Phalaris plants 3–4 years old, and therefore the majority of these crops were not fully mature. Other studies have shown differences between mineral composition of young and mature plants (Monti et al., 2008). Further long-term studies are therefore required.


This research was funded by the European Regional Development Fund. Thanks are due to ADAS Pwllpeiran, Paul Radcliffe and the Llysdinam Charitable Trust for access to sites. Thanks also to the Cardiff University professional training year students who helped with the field work: Amy Tibble, Robert Firmin, Fiona Ross, Eleanor Sherrard-Smith, Kate Davidson, Claire James, James Chen, Annabelle Phillips, Victoria Tombs, Emily Howe, Kate McCrum and also to Alun Martin at ADAS Pwllpeiran.