LAWR

October 2014 Local Authority Waste & Recycling 27 MAPPING THE MRF from 2,000 households which will rep- resent many millions of bottles, trays and tubs. The magic number When drawing the sample from that huge pile, you need to understand that you must take the right number of randomly selected items. This 'N' is the number which drives all of the subsequent statistics derived from this particular sample. Because it's N that dictates what is a sensible sample size, N also sets the confidence level of how well your small sample truly represents the over- all 'population' – and crucially, it's N which sets the error band that you can expect either side of your final result. It is worrying that on so many occa- sions we see people writing down a sample test result with 3 or 4 deci- mal places, eg: 'plastic = 76.2345%'. It would be much more statistically rigor- ous and realistic to see reported sam- pling results expressed as 'the plastic composition is probably 80% plus or minus 5% with a 95% confidence level based on the measured sample size of N number of items'. Looking ahead as to how the system will turn out, I think the bigger players will adopt rigorously robust and statis- tically-meaningful sampling systems because they have the scale and the skills in-house to deliver good results. I really hope that we see results reported with error bands and confidence levels on a regular basis. However, I'm concerned that we may see some numbers reported from smaller players who simply weigh 20kg of material every two weeks and stick a number in a box which won't have much meaning. So I'm hoping that it really will lead to some realism, more science and greater statistical rigour in the UK marketplace that will increase the efficiency of the trading which takes place in our sector and highlight the value of quality all along the line. Not all facilities are affected by the new regulations. Those accepting less than 1,000 tonnes of mixed waste a year are exempt, as are councils who source-separate their materials. I think that's fair; you have to draw the line somewhere. Many smaller facilities tend not to be involved in the big stream of house- hold waste. They're usually smaller skip operators who are dealing with skips from households and industrial waste, and I think 1,000 tonnes is a reasonable cut-off point. It's only 20 tonnes a week, which could be as lit- tle as 10 skip pick-ups from outside houses, so they're a not a big operator. Any MRF operators or local authori- ties professionally managing an opera- tion that's above the 1,000-tonne threshold should already be fully pre- pared for the new system. Let's face it; in recent months it would have been difficult to miss all the debate about the MRF CoP! Additionally, they should be contact- ing the experts who understand the sta- tistics, techniques and practicalities of good sampling practice, then installing what equipment they need and training their staff to do the job correctly. What next? They'll be looking at major benefits for their business as they'll have better knowledge of what's going on in their operations, while the amount of additional work is compara- tively minimal. However the benefits in terms of value that they sell to their custom- ers will be greater because they can offer known and measured products; as opposed to a 'buyer beware, come and have a look in the yard and tell me what the price is' basis. Keith Freegard is director at resource recovery specialist Axion Consulting. ► The first reporting period starts this month. ► Specifically, one sample must be taken for every 160 tonnes of mixed waste material received at the materials facility from each supplier during each reporting period before 1st October 2016, rising to one sample per every 125 tonnes on or after that date. Material specifications of each sample must also be listed. ►Samples must also be taken of the output streams for target materials when they reach a certain level of throughput. Before 1 October 2016, the tonnages are as follows: • Glass: 50 tonnes • Paper: 80 tonnes • Metal: 20 tonnes • Plastic: 20 tonnes ► After this date, paper samples drop to once every 60 tonnes and plastic to 15 tonnes, while metal and glass sampling obligations remain the same. ► The total weight of all the samples taken must provide an average weight of 60 kilogrammes or more per sample (and no less than 55 kilogrammes). Source: Defra How the sampling regime works

• Cover