Silk Revisited

Since this pro bono article was published in Apparel Insider Issue 14, July 2020, the Sustainable Apparel Coalition (SAC) and Higg Co have amended the Higg MSI for silk.

It has increased from 681/kilo, to 1086/kilo.

Neither the SAC nor Higg have offered any explanation for this, and the stated source, remains unchanged:

"Meta Information: Primary Source Quantis. World Apparel and Footwear Life Cycle Assessment Database (WALDB), version 2.0, 2018. Last Publication or Review date 2018 Geography China/India Data Generator Quantis"

In issue 11, of Apparel Insider, I wrote a brief piece questioning the Sustainable Apparel Coalition's (SAC) Higg index and its evaluation of silk as the world’s least sustainable fiber. The Higg claims that silk has a total MSI impact of 681 per kilo, whilst fast fashion favourite, polyester, has a total impact of only 44 per kilo. That is a massive difference.

Brands using the Higg - along with The Global Fashion Agenda and a host of other initiatives - are telling us that if ASOS for example, dumps a kilo of silk from its production line and brings in 10 kilos of polyester instead, it is saving the planet.

Seriously?

By pure coincidence no doubt, since my piece was published a number of articles have appeared in magazines and blog posts, supporting and purportedly substantiating the Higg’s position.

Recent Claims and assertions:

So, let’s unpack some of the recent claims and assertions and see where it takes us. In doing this it is important to remember:

1) Silk is light and valuable - if you get more shirts from a kilo of silk than from a kilo of cotton then simply comparing impact per kilo would be misleading.

2) That clothes are not kleenex. You use them more than once. If a silk shirt has a production impact of 681, but is worn 50 or 100 times, it is still far more sustainable than a polyester shirt with an impact of 44 that is worn 2 or 7 times - shedding polluting fibers to air and water all the while, and clogging landfill for decades thereafter. Simply looking at impact up to the factory gate, as the Higg does, tells you virtually nothing. 

Why do neither the SAC nor its supporting brands and initiatives ever make that clear?

3) There are many kinds of silk. At the risk of oversimplifying, silk as commonly known comes from Bombyx Mori, but there are two main types: bivoltine and multivoltine (and yes, there are crossbreeds). The former produces better silk, but prefers a temperate climate; the latter is more hardy and so happy in a tropical climate, but doesn’t produce the kind of silk generally used by the global apparel industry. India’s silk industry is almost entirely polyvoltine, China’s, bivoltine.

Keeping all that in mind then, what has this recent blossoming of articles told us - and is any of  it true?

The first thing we discovered is that the source of the Higg values is a 2014 study by a team from the University of Oxford: “Life cycle assessment of Indian silk” by Miguel F. Astudillo, Gunnar Thalwitz, Fritz Vollrath. The study is behind a paywall, but the lead author was so kind as to both send it to me, and to answer my questions and queries as best he could recollect. That said, all interpretations and conclusions made here are my own.

Besides the purported source, the next thing we were told, is that where silk causes the most environmental damage is in its energy usage. We were told this is because silk farms are kept at 65 degrees humidity and temperature, and it takes a lot of energy to cool a warehouse to 65 degrees in South and Southeast Asia.

Given silk originated in S and SE Asia, thousands of years ago when there was no air conditioning, that seems a very odd claim. And if we look at the LCA Higg are using, it actually asserts something rather different: “Most GWP100 of silk is related to high levels of fertilizer and FYM use.” (FYM=farmyard manure GWP100 = global warming potential over 100 years)

So the energy value does NOT relate to air conditioning, which - I checked with the authors - the 100 farmers studied in 2006, did not in fact use “Farmers were not using air conditioning. Silkworm rearing is often done in quite rudimentary settings.”

Hmm. 

Well where did that 65 degrees come from then? When Apparel Insider enquired further, the figure was attributed to Higg CEO Jason Kibbey. The lead author of the Oxford study that the Higg is claiming to use, on the other hand, recollects that silkworms were kept at a temperature of “about 30 ºC or lower”.

If you search online, this - or some variant of same -  is what you will find:

“The optimum temperature for normal growth of silkworms is between 20°C and 28°C and the desirable temperature for maximum productivity ranges from 23°C to 28°C. Temperature above 30°C directly affects the health of the worm. If the temperature is below 20°C all the physiological activities are retarded, especially in early instars; as a result, worms become too weak and susceptible to various diseases”.

Well in that case, since 65F equates to a little over 18C, we have to conclude that one of the recent claims put forward to substantiate the Higg, is completely false - silk farms are definitely not cooled to 65F.

A second claim made in the recent articles is that industrially produced silk has a surprisingly high environmental impact, but as we have already seen, the 2014 LCA doesn’t cover industrially produced silk, it covers production by smallholder farmers in rural India. Indeed, I can find no publicly available LCA of industrial silk production and none of those articles actually linked to one. A second of the recent claims then, can be thrown out. It’s speculation and completely unsubstantiated.

A third reason cited for silk’s poor environmental rap is its water consumption - indeed looking at the Higg Index itself, it is not in fact energy usage that causes the most damage, but Water Scarcity. This is the single most important impact at 367.7/kilo.

Does the 2014 Oxford study at least substantiate this claim? Well yes - and no.

The University of Oxford Silk LCA:

The Oxford LCA actually computes two different sets of impact values - one for production according to Recommended Practices (obtained from a 2013 publication by the Government of Andhra Pradesh). The other, obtained from the actual practices of 100 bivoltine silk farmers in Dharmapuri, Tamil Nadu - as observed in 2006. 

Which impact valuation the SAC are using has not been made clear, but I am guessing the values for the actual farm practices in Dharmapuri, as they are significantly higher.

So what kind of levels of irrigation were measured in 2006, for these 100 Tamil Nadu silk farmers? Unfortunately, the answer is none whatsoever. Their irrigation was not measured at all.

For the LCA, irrigation had to be estimated. To do this, the authors took the crop water requirement of 16,000 m3 /ha/y  for mulberry cultivation on loamy soils (as provided in a 2003 study) and deducted average monthly rainfall: “Blue water withdrawal for irrigation was estimated as the difference between crop water requirements and effective rainfall in Bangalore..., where monthly rainfall rarely exceeds the water requirements.”

I asked the Oxford authors why Bangalore, Karnataka, rainfall was used when the farmers studied were in Dharmapuri, Tamil Nadu. I also asked why, when the 100 farmers deviated from recommended practices in every other area, it was assumed that they adhered to recommendations for irrigation.

Given that the farmers did not produce silk in the summer (when the mulberry leaves could only be sold as fodder of little worth); the 100 farmers were obtaining only 70% of the expected yield of Mulberry leaf (tonnes/ha); and were only using 78% of mulberry for rearing (the recommended practice is 95%). I posited that it was very possible they were also using only, say, 75% of the recommended irrigation (keeping the trees alive in the months when silk is not produced, rather than maximising leaf yield). 

The lead author replied: “It is possible that farmers also irrigated less than recommended. It is also possible that they irrigated more. The article we used as a reference did not have info on irrigation, so we assumed they irrigated following guidelines”

Quite. 

So does the Oxford study have anything else to say that would impact the SAC’s contention that for mulberry cultivation “Common practices require more than 8,500 cubic meters of water per hectare per year and more than 9,000 cubic meters during dry seasons. (Astudillo et al. [2014] and Huo [2017])”

Yes.

But before discussing this it's important to remember that not all Indian silk production is irrigated - some is entirely rainfed, how much of that is bivoltine, I cannot say, but it does mean 8,500 m3 is not the average, it’s the maximum. Second, the Tamil Nadu government has an active program promoting rural rainwater harvesting - including conversion of defunct borewells. Does this now include the silk farmers of Dharmapuri? I don’t know.

So back to the report. The Discussion section highlights a number of issues that contribute to the high impact values found, but none of these appear to have been taken into account in the Higg calculations.

Issues and considerations:

As the Oxford authors note:

1. Co-Products:

Because the co-products are insufficiently valorised in India, the study outcome is almost total attribution of impact to reeled silk. Since Pupae and sericin constitute over 50% of dry weight of final output, apportionment by value means that in China, where the uses of silk co products are multiple, the impact attributed to silk - including that of irrigation - would be considerably lower.

2. Solar Power:

Solar dryers for silkworm cocoons have been developed, reducing electricity requirements ten-fold compared with the electric dryers employed in the study. Since 2014, India has had a major push towards rural off grid electrification; if the majority of farmers are now using solar pumps, irrigation impact on emissions could also have been significantly reduced.

3. Drip irrigation:

Considerable energy and water can be saved using drip irrigation; the study mentions water savings of 66% can be obtained without compromising mulberry yields.

A quick search of recent articles on Indian sericulture reveals that such irrigation is being actively promoted, and farmers discouraged from using flood irrigation - something many silk farmers have apparently had to give up anyway, due to a lack of available water in some areas.

4. India doesn’t actually matter:

Most importantly of all, as the 2014 Oxford study states: “It also bears repeating that gradable bivoltine silk, representative of internationally traded silk, constitutes only a small proportion of total Indian silk output.”

Virtually none of the silk consumed by the apparel sector of the global North actually comes from India. So why is the SAC’s valuation based on an LCA of Indian production?

Where did China go?

To be fair to the Higg, they claim that the huge negative impact of silk production does not just come from the Oxford study: “Common practices require more than 8,500 cubic meters of water per hectare per year and more than 9,000 cubic meters during dry seasons. (Astudillo et al. [2014] and Huo [2017])” So they are suggesting that their values are based on Chinese data as well - which, as we have just established, is the only data that really matters for the average production impact of silk consumed by the global apparel sector.

The only problem is that Huo 2017, does not exist.

There is a 2000 FAO study "Mulberry cultivation and utilization in China" by Yongkang Huo.  This is the only Huo silk study I could find. When Apparel Insider attempted to confirm that this was the study being used, the SAC did not reply. So I checked with the Oxford group, and that is the only Huo study their lead author was aware of.

And what sort of study is it?

It is not an LCA. First, the Huo report lists the eight Chinese provinces where silk is grown, with Geographical position, Climatic features, and Main mulberry varieties cultivated. It then discusses the Taxonomy by which the genus Morus is classified into 14 species and 1 variety. 

The study then moves on to the more than 1,000 cultivated varieties of mulberry in China, listing average yield, fertiliser recommendations etc for the 27 major varieties. It ends with some general cultivation methods - propagation, planting, field management - and finally, “other traditional uses” including tea and edible fungi.

Were this study  to be current, which it is not, it contains none of the data required to produce a robust LCA. For the Higg to suggest that they evaluated current Chinese silk production on this basis would be laughable, were it not such a massive disservice and injustice to some of the poorest people on the planet.

Conclusion:

If you want to make comparative sustainability assertions it is your responsibility to obtain the substantiating data. Brands do not get to unilaterally decide to extrapolate directly from a single, unrepresentative study, because it fits their agenda. If the SAC does not have a study for China, then we cannot accept any of their assertions about silk until they do.

Huge sums of money are spent in the sustainability sector. A quick check of just one initiative - the C&A Foundation (C&AF - now Laudes Foundation), shows they have given €2.6 million to Fashion Revolution; €2 million to the Better Cotton Initiative; €7.5 million to CottonConnect; €1.1 million to the Organic Cotton Accelerator; €1.5 million to the Sustainable Apparel Coalition itself and scores of other grants to various initiatives, totalling tens of millions of Euros. 

By comparison, the robust and comprehensive 2018 assessment of the socio-economic impact of three cotton cultivation systems in India - conducted by AIR USA - cost C&AF a paltry €240,000. 

To suggest that there is no money available to conduct robust, independent LCAs of all farmed fibers is contradicted by the evidence.

Anyone serious about actually producing more sustainably, rather than merely claiming to do so, would view such studies - along with companion SEIAs/randomised controlled trials etc - as an absolute priority.

What then should we conclude about the intentions of the SAC and its members?

This article was published in Apparel Insider Issue 14, July 2020