S1

—Appendix. Supplementary Information—

In situ Degradation of Biodegradable Plastic Mulch

Films in Compost and Agricultural Soils

Henry Y. Sintima,b, Andy I. Barya,b, Douglas G. Hayesc, Larry C. Wadsworthc, Marife

B. Anunciadoc, Marie E. Englishc, Sreejata Bandopadhyayc, Sean M. Schaefferc,

Jennifer M. DeBruync, Carol A. Milesd, John P. Reganoldb, and Markus Flurya,b,∗

aDepartment of Crop & Soil Sciences, Washington State University, Puyallup, WA 98371

bDepartment of Crop & Soil Sciences, Washington State University, Pullman, WA 99164

cDepartment of Biosystems Engineering and Soil Science, University of Tennessee,

Knoxville, TN, 37996

dDepartment of Horticulture, Washington State University, WSU Mount Vernon,

Northwestern Washington Research & Extension Center, Mount Vernon, WA 98273

Supplementary information includes details on ASTM D5988 test, mulch and soil character-

ization, meshbag study.

S2

S1 ASTM D5988 Biodegradation Test

Field-weathered BioAgri and PLA/PHA were retrieved from Knoxville field plots after the

pumpkin cropping season in 2017 (see below). Unused cellulosic paper mulch was used as

control. The mulch films were cut into 1 cm × 1 cm pieces (0.21 to 0.25 g) and added to

the soil (50 g) with a C:N ratio of 10:1 (g/g). The mixture was then placed into sample

jars and incubated at 27 oC following the ASTM D5988 standardized test protocol (ASTM

International, 2012). The soil, a Shady-Whitwell complex soil, was sieved (2 mm) and plant

debris were removed. Air samples were taken from the headspace of the jars, and CO2 evolu-

tion was measured using a CO2 infrared gas analyzer (LICOR 820,Lincoln, NE). The percent

biodegradation was calculated based on ASTM D-5988. Moles of CO2 were determined and

converted to microgram of carbon. Biodegradation was determined by the ratio of cumulative

carbon produced over the amount of original carbon from mulches added to the sample jar

(Figure S1).

S2 Physicochemical Characterization of Mulches

Analyses were performed on the initial mulches (as received from manufacturers) and after

deployment in the field. Paper mulch underwent complete macroscopic disintegration after

field deployment at Knoxville, and so no samples were remaining for analyses. The field-

weathered mulches were cleaned by gentle dry brushing to remove adhering soil and water

before the analyses. Details of the cleaning procedure and analyses used to characterize the

physicochemical compositions are reported elsewhere (Hayes et al., 2017). Below, we provide

a brief description of the methods.

S3

S21. Physical Analyses

Weight was determined gravimetrically on the mulch samples, with dimensions of 17.78 cm in

the machine direction and 15.24 cm in the cross-machine direction, and the same specimens

were used for thickness measurements with a TMI 49-70 Series Micrometer (Testing Machines

Inc., Amityville, NY). Percent elongation at maximum tensile stress in the machine direction

was determined according to ASTM D5035 (Model 5567, Instron, Norwood, MA) and a 10 kN

load cell. An initial gauge length between the clamps of the tensile tester was 2.54 cm instead

of the recommended 7.72 cm, because most of the field-weathered mulches were smaller

than the recommended specimen length of 15.24 cm. Water contact angle was measured

by the sessile drop method using a manual goniometer (model 147 50-00-115, Rame-Hart

Instrument Co., Netcong, NJ) at 20oC. Glass transition temperature (Tg) and melting point

temperature (Tm) were determined via differential scanning calorimetry; 3 to 7 mg samples

were analyzed using a model Q 2000 calorimeter from TA Instruments (New Castle, DE),

except the polyethylene mulch where the Tg was not measured, but values were taken from

the literature. Heating-cooling cycle for the differential scanning calorimetry was as follows:

heating at 10oC min−1 from 40oC to 200oC, and the temperature held constant at 200oC for

5 min, followed by cooling at 10oC min−1 until reaching −50oC, and the temperature held

constant at −50oC for 5 min. A second heating-cooling cycle was performed using the same

heating rate as the first cycle, except for polyethylene where a single heating-cooling cycle

was used.

S4

S22. Chemical and Molecular Analyses

Total carbon content (%C) and δ13C in the mulch samples were determined using a cavity

ring-down spectrometer (CRDS, Picarro G2121-i, Picarro Inc., Santa Clara, CA) coupled to

a combustion module (CM-CRDS, Costech Analytical Tech Inc., Valencia, CA). Internal lab-

oratory standards were used for all isotopic analyses, which were calibrated against certified

standards (USGS 41 and USGS 40, L-glutamic acid). Gel permeation chromatography was

used to determine the weight-averaged molecular weight of chloroform-soluble components

from the mulches. Only BioAgri, Organix, and PLA/PHA samples were analyzed.

S3 Effect of Meshbags on Mulch Degradation

Meshbags will likely affect mulch degradation because they hinder mass and energy transfer

between the mulches and the soil outside of the bags. The larger the meshbag opening, the

less is the effect on mulch degradation; however, large openings also allow non-degraded mulch

pieces to fall out, making quantifications difficult. In order to determine which meshbag size

would be suitable, we setup a pilot study in October 2014. We composted a biodegradable

plastic mulch (Bio360, manufactured by Dubois Agrinovation, Quebec, Canada), a PLA

fabric, and a paper mulch enclosed in 250-µm and 1-mm nylon meshbags. Mulch was placed

into meshbags (no compost was added to the meshbags so that we could retrieve the mulch

pieces without disturbance), and the meshbags were then placed into compost. The compost

was an aerated static pile containing broiler litter (28% vol.), dairy manure solids (28%), fish

carcasses (2%), bedding (14%), and yard wastes (28%). After 18 weeks of composting, we

could not visually detect mulch samples in either meshbag, except for black staining (more

conspicuous on 250-µm nylon meshbag) that was observed on the meshbags that contained the

S5

Bio360 (Figure S2). Thus, we concluded that the 250-µm opening was large enough to allow

access of bacteria and fungi to degrade the mulches during our composting study. Nonetheless,

the meshbags likely will have slowed down the degradation of the mulches in compost and soil;

however, without the use of meshbags it would be impossible to quantitatively recover buried

mulch samples. While the kinetics of the degradation may be affected by the meshbags,

the overall relative differences in the degradation among the mulches should not drastically

change. In addition, the use of paper mulch as a positive control treatment in this study will

enable us to determine the extent to which meshbags affect mulch degradation.

S4 Measurement of Soil Properties

Soil water content and temperature were monitored with sensors installed at 10-cm depth

(5TM sensor logged with EM50G data logger, Decagon Devices Inc., Pullman, WA). Other

soil properties were measured in September 2015, prior to mulch burial in soil. Soils were

sampled in the top 0 to 12 cm depth (composite of 15 random samples) with a soil core

sampler, air dried in the laboratory, and sieved through a 2-mm sieve. The soil samples

were then analyzed for soil pH, organic matter, nitrate-N, and available phosphorus by a

commercial laboratory (American Agricultural Laboratory, McCook, NE) following standard

soil testing procedures (NCERA, 2015). The composite soil samples were also analyzed for

soil respiration using the soil CO2-C burst Solvita test method (Woods End Laboratories,

2016).

Four enzymes were assayed: β-glucosidase, β-D-cellubosidase , β-xylosidase, and N-

acetyl-β-glucosaminidase, using 4-MUB-β-D-glucopyranoside, 4-MUB-β-D-cellobioside, 4-MUB-

β-D-xylopyranoside, and 4-MUB-N-acetyl-β-D-glucosaminide substrates, respectively. For

S6

enzyme analyses, soil was mixed with a sodium acetate trihydrate buffer, adjusting the pH

closely with respective soil pH values measured at the two sites. An aliquot of the soil slurry

(800 µL) was pipetted into 96-DeepWell plates, and 200 µL of appropriate standards and

substrates were added to the soil slurries. Separate plates were prepared using standard

curves of 4-methylumbelliferone and 7-amino-4-methylcoumarin synthetic fluorescent indi-

cators. The plates were sealed, inverted to mix the contents, and then incubated at room

temperature for 3 hours. The substrates and standard plates were centrifuged at 1500 rpm,

and the supernatants pipetted into black 96 well plates. Fluorescence microplate enzyme

assay was then performed using a BioTekTM plate reader (Winooski, Vermont) at excitation

wavelength of 365 nm and emission wavelength of 450 nm.

DNA from soil samples (0.25 g) was extracted with the MoBioTM PowerLyzerTM Power-

Soil DNA isolation kit with inhibitor removal technology. Extracted DNA was quantified with

the Quant-ItTM PicoGreenTM dsDNA Quantification Kit (ThermoFisher Scientific). Bacte-

rial and fungal abundances (16S rRNA and ITS gene copies, respectively) were quantified

from soil DNA samples using established protocols (FemtoTM Bacterial DNA quantification

kit and FemtoTM Fungal DNA quantification kit, Zymo Research). qPCR was done with a

CFX Connect Real-Time PCR Detection System (BioRad). All samples were analyzed in

triplicate.

DNA samples were shipped frozen to the Genomic Services Laboratory (GSL) at Hud-

son Alpha (Huntsville, AL) for 16S rRNA amplicon sequencing. The V4 region of the 16S

rRNA gene was amplified using primers 515F (GTGCCAAGCAGCCGCGGTAA) and 806R

(GGACTACHVGGGTWTCTAAT), and the first PCR was run with V4 amplicon primers,

Kapa HiFi master mix, and 20 cycles of PCR. PCR products were purified and stored at

S7

−20oC. The PCR indexing was later completed for the 16S (V4) amplicon batch. Products

were indexed using GSL3.7/PE1 primers, Kapa HiFi master mix, and 12 cycles of PCR. Final

libraries were quantified using Pico Green.

Raw sequence data was used to trim off primers, make contigs, remove sequences with

ambiguous bases and long homopolymers using mothur v.1.39.5 following the MiSeq SOP.

Before aligning to the reference database (SILVA release 102), unique sequences were identi-

fied. After alignment to SILVA database, sequences were filtered to remove overhangs at both

ends, and sequences de-noised by pre-clustering sequences with up to 2 nucleotide differences.

Chimeras were removed using the VSEARCH algorithm. All sequences including 18S rRNA

gene fragments and 16S rRNA from Archaea, chloroplasts, and mitochondria were classified

using the Bayesian classifier against the mothur-formatted version of the RDP PDS training

set (v.9) with a bootstrap value of >80% cut off and the general approach to using 80%

confidence in bootstrap values for phylogenetics. Following this step, untargeted (i.e., non-

bacterial) sequences classified as Eukaryota and Arachaeota were removed, and sequences

were binned into phylotypes according to their taxonomic classification at the genus level.

Beta diversity was computed using Bray-Curtis distances of microbial community com-

position (vegan package v 2.4-3 in R). Significant differences between bacterial community

composition at the two locations for Fall 2015 were analyzed by permutational multivariate

analysis of variance (ADONIS function in R) based on the Bray-Curtis dissimilarity matrix.

All libraries were scaled to even depth (minimum sample read count, i.e., smallest library

size of 34,328) before analysis was performed. Alpha diversity was computed by subsam-

pling the libraries to the minimum number of reads (34,328), with replacement to estimate

species abundance of the real population by normalizing sampling effort. The subsampling

S8

was repeated 100 times and the diversity estimates from each trial were averaged. The R phy-

loseq package (estimate richness function) was used to calculate richness (number of observed

OTUs) and Inverse Simpson index (for diversity).

S5 Estimation of Mulch Degradation in Soil

Upon retrieval from the soil, the meshbags containing the mulches were cleaned gently with a

soft brush and a slightly moistened terry cloth to remove adhering compost or soil, and then

air-dried. The meshbags were then cut open and the mulch pieces transferred onto a white

surface. The mulch fragments were spread uniformly on the white surface and photographed.

ImageJ software was used to digitize the photographs and measure the total area. The

total area was then expressed as percent or fraction of the initial surface area of the buried

mulch samples. The extent of mulch degradation (EoD) in equation (2) was evaluated by

numerically integrating the measured function f(T ) by the trapezoidal rule.

A 2nd-order polynomial model was used to estimate the degradation of BioAgri, Na-

turecycle, Organix, and PLA/PHA after soil incubation in Knoxville and Mount Vernon.

Mulch degradation was modeled as a function of thermal time, with 0oC base temperature,

and calendar time, separately for the two locations. Model performance was assessed by the

mean absolute error (MAE), normalized root mean squared error (NRMSE; normalized to

the range of observed data), coefficient of determination (R2), and index of agreement (d):

MAE =1

n

n∑i=1

|Pi −Mi| (S1)

NRMSE = 100

(1n

∑ni=1(Pi −Mi)

2

Mmax,i −Mmin,i

)(S2)

S9

R2 =

∑ni=1(Mi −Mave)(Pi − Pave)(√∑n

i=1(Mi −Mave)2) (√∑n

i=1(Pi − Pave)2)2

(S3)

d = 1−( ∑n

i=1(Pi −Mi)2∑n

i=1(|Pi −Mave|+ |Mi −Mave|)2

)(S4)

Mi and Pi are the measured and predicted values, respectively; Mave and Pave are the

average of the measured and predicted values, respectively, and n is the number of data;

MAE and NMSE ranges from 0 to +∞, where 0 is a perfect fit. R2 and d ranges from 0 to 1,

where 1 indicates a perfect fit.

References

ASTM International (2012). Standard Test Method for Determining Aerobic Biodegradation

of Plastic Materials in Soil, D5988–12. ASTM International, West Conshohocken.

Hayes, D. G., Wadsworth, L. C., Sintim, H. Y., Flury, M., English, M., Schaeffer, S. and

Saxton, A. M. (2017). Effect of diverse weathering conditions on the physicochemical

properties of biodegradable plastic mulches. Polymer Testing 62, 454–467.

Khonakdar, H. A., Jafari, S. H. and Hassler, R. (2007). Glass-transition-temperature depres-

sion in chemically crosslinked low-density polyethylene and high-density polyethylene

and their blends with ethylene vinyl acetate copolymer. J. Appl. Polym. Sci. 104, 1654–

1660.

Laredo, E., Suarez, N., Bello, A. and Marquez, L. (1996). The glass transition in low linear

S10

density polyethylene determined by thermally stimulated depolarization currents. J.

Appl. Polym. 34, 641–648.

NCERA (2015). Recommended chemical soil test procedures for the North Central Region.

North Central Regional Res. Publ. No. 221 ed. North Central Extension and Research

Activity, Madison, WI.

Trautmann, N. M. and Krasny, M. E. (1998). Composting in the classroom: scientific inquiry

for high school students. Chapter 1; The science of composting. Kendall/Hunt Publishing

Company, Dubuque, IA, USA.

Woods End Laboratories (2016). Soil CO2-burst official Solvita instructions. Woods End

Laboratories, Inc., Mt Vernon, ME.

S11

Tab

leS

1.P

hysi

coch

emic

al

pro

per

ties

ofth

ep

last

icm

ulc

hes

bef

ore

and

afte

rd

eplo

ym

ent

inth

efi

eld

du

rin

g20

15.

Val

ues

taken

from

Hay

eset

al.

(2017

),ex

cep

tw

hen

not

edot

her

wis

e.V

alu

esre

pre

sent

the

mea

n±

stan

dar

dd

evia

tion

(n=

4).

Mulc

hes

Wei

ght

Thic

knes

sE

longati

on

Mole

cula

rw

eight

Conta

ctangle

Tota

lca

rbon

δ13C

Tg

Tm

(gm−2)

(µm

)(%

)(k

Da)

(o)†

(%)

(h)

(oC

)(o

C)

Init

ial

Bio

Agri

23.1

±0.6

26.0

±2.0

260±

35

246±

12.8

88±

657.6

±1.8

−26.2

±0.2

−30.7

±0.5

93.5

±2.0

Natu

recy

cle

22.7

±1.6

48.0

±2.0

213±

45

na

69±

554.8

±0.6

−31.5

±0.2

−31.2

±0.4

105±

0.2

Org

anix

19.2

±0.7

20.0

±2.0

273±

29

251±

30.7

86±

451.4

±0.1

−29.8

±0.3

−30.1

±1.5

120±

0.6

PL

A/P

HA

26.2

±0.6

33.0

±2.0

247±

17

267±

7.3

68±

347.5

±0.2

−13.0

±0.0

49.2

±1.2

154±

0.0

Pap

er109±

1.2

479±

18

6.4

±0.6

na

<10

46.0

±1.8

−25.8

±0.2

na

na

Poly

ethyle

ne

22.5

±0.5

47.0

±2.0

578±

32

na

79±

282.9

±1.2

−30.0

±0.1

−120‡

110±

0.0

Knox

ville

(fiel

d-w

eath

ered

)

Bio

Agri

25.4

±1.9

78.0

±6.0

14.2

±7.1

185±

6.1

41±

346.0

±9.5

−25.8

±0.2

−28.9

±1.0

108.0

±1.7

Natu

recy

cle

38.9

±2.8

132±

2.0

6.1

±0.6

na

37±

751.6

±0.8

−31.3

±0.1

na

na

Org

anix

20.7

±0.4

74.0

±8.0

11.3

±1.1

211±

3.9

47±

248.2

±1.5

−29.1

±0.1

−29.4

±0.2

110±

2.5

PL

A/P

HA

28.9

±0.7

58.0

±4.0

5.3

±0.4

226±

4.5

52±

643.5

±1.2

−12.5

±0.1

49.7

±2.6

155±

0.1

Pap

erna

na

na

na

na

na

na

na

na

Poly

ethyle

ne

24.3

±0.5

59.0

±4.0

411±

35

na

53±

777.2

±0.5

−30.1

±0.1

na

110±

0.1

Mount

Ver

non

(fiel

d-w

eath

ered

)

Bio

Agri

23.0

±0.8

38.0

±4.0

28.4

±5.4

297±

6.6

40±

2na

na

−30.0

±1.3

98.7

±5.3

Natu

recy

cle

28.4

±1.2

58.0

±2.0

7.9

±0.9

na

56±

3na

na

na

na

Org

anix

19.5

±0.4

31.0

±2.0

79.5

±20.4

265±

11

35±

4na

na

−29.2

±0.4

110±

1.7

PL

A/P

HA

26.6

±0.3

37.0

±2.0

7.2

±0.3

218±

6.5

59±

2na

na

48.3

±2.8

154±

0.1

Pap

er111±

0.4

522±

20

5.6

±0.4

na

<10

na

na

na

na

Poly

ethyle

ne

26.3

±0.4

50.0

±4.0

378±

35

na

78±

3na

na

na

110±

0.1

Tg:

Gla

sstr

an

siti

on

tem

per

atu

re;

Tm

:M

elti

ng

poin

tte

mp

eratu

re.

† :M

easu

red

at

20oC

.‡

Valu

eta

ken

from

Lare

do

etal.(L

are

do

etal.,

1996)

an

dK

hon

akd

ar

etal.(K

hon

akd

ar

etal.,

2007).

na:

Data

not

availab

le.

Th

eT

gan

dT

mof

PL

A/P

HA

refl

ects

that

of

PL

A,

wh

erea

sth

at

of

Bio

Agri

,N

atu

recy

cle

an

dO

rgan

ixre

flec

tsP

BA

T.

Mole

cula

rw

eight

refl

ects

chlo

rofo

rm-s

olu

ble

com

pon

ents

from

the

mu

lch

es.

S12

Tab

leS

2.B

acte

rial

and

fun

gal

ab

un

dan

cefo

rd

iffer

ent

mu

lch

trea

tmen

tsfo

rF

all

2015

and

Fal

l20

18.

Ab

un

dan

ces

are

inlo

g10

gen

eco

pie

sg−

1d

ryso

il.

Tre

atm

ent

Knox

ville

Mount

Ver

non

Fall

2015

Fall

2018

Fall

2015

Fall

2018

16S

ITS

16S

ITS

16S

ITS

16S

ITS

Bio

Agri

9.3

96±

0.1

60

8.5

85±

0.1

41

9.1

43±

0.3

59

8.3

72±

0.5

38

9.2

91±

0.4

30

8.152±

0.397

9.7

29±

0.2

18

9.777±

0.312

Natu

recy

cle

9.5

07±

0.0

45

8.7

36±

0.0

79

9.1

75±

0.2

61

8.2

05±

0.3

25

9.6

42±

0.2

22

8.481±

0.361

9.8

08±

0.1

07

9.530±

0.094

No

Mulc

h9.5

11±

0.3

01

8.6

94±

0.3

69

9.2

18±

0.2

25

8.3

90±

0.5

29

9.683±

0.128

8.867±

0.201

9.948±

0.159

9.658±

0.103

Org

anix

9.6

86±

0.3

62

8.6

29±

0.3

15

9.3

21±

0.2

44

8.9

09±

0.2

87

9.370±

0.049

8.469±

0.148

9.799±

0.045

9.521±

0.088

PL

A/P

HA

9.3

15±

0.2

54

8.4

83±

0.2

42

9.1

17±

0.5

17

8.6

80±

0.9

46

9.405±

0.171

8.301±

0.190

9.784±

0.157

9.526±

0.111

Poly

ethyle

ne

9.2

51±

0.2

10

8.213±

0.311

9.2

44±

0.0

92

8.841±

0.419

9.274±

0.219

8.280±

0.146

9.812±

0.079

9.625±

0.073

Pap

er9.3

21±

0.2

31

8.4

76±

0.3

42

9.1

71±

0.5

48

8.4

17±

0.8

19

9.565±

0.222

8.453±

0.262

9.677±

0.210

9.329±

0.174

16S:

bact

eria

lm

ark

ergen

e

ITS:

fungal

mark

ergen

e

Num

ber

sin

bold

face

indic

ate

signifi

cant

diff

eren

ces

atP

=0.0

5of

bact

eria

land

fungal

gen

eco

pie

sb

etw

een

Fall

2015

and

Fall

2018

S13

Table S3. Chemical and biological properties of the soil at Knoxville and Mount Vernon,before meshbag burial in Fall 2015.

Soil parameters Units Sites

Knoxville Mount Vernon

Soil pH 5.70 ± 0.39 5.82 ± 0.23

Nitrate-N mg kg−1 10.7 ± 3.9 17.0 ± 15.9

Available P mg kg−1 79.9 ± 28.0 91.4 ± 10.4

Organic matter g kg−1 15.2 ± 1.9 23.4 ± 1.3

CO2-C mg kg−1 8.52 ± 3.30 40.9 ± 15.1

β-glucosidase µmol activity hr−1 g−1 soil 40.0 ± 14.0 47.2 ± 14.7

β-D cellobiosidase µmol activity hr−1 g−1 soil 8.58 ± 5.03 13.2 ± 4.4

β-xylosidase µmol activity hr−1 g−1 soil 1.06 ± 1.53 2.41 ± 1.55

N-acetyl β glucosaminidase µmol activity hr−1 g−1 soil 8.43 ± 5.75 15.2 ± 4.7

Values represent the mean ± standard deviation (n = 24).

S14

Tab

leS

4.B

act

eria

lan

dfu

ngal

abu

nd

an

cean

dd

iver

sity

and

rich

nes

sin

the

soil

atK

nox

vil

lean

dM

ount

Ver

non

inF

all

2015

.A

bu

nd

ance

sare

per

dry

soil

mas

s.

Loca

tion

Bac

teri

alab

un

dan

ceF

un

gal

ab

un

dan

ceD

iver

sity

Ric

hn

ess

log10

(16S

gen

eco

pie

sg−1

soil

)lo

g10

(IT

Sgen

eco

pie

sg−1

soil

)(I

nve

rse

Sim

pso

nin

dex

)(N

rof

ob

serv

edO

TU

s)

Kn

oxvil

le9.

41±

0.16

8.5

2±

0.1

87.95±0.27

271±10

Mou

nt

Ver

non

9.43±

0.15

8.3

6±

0.1

38.71±0.37

259±7

Nu

mb

ers

inb

old

face

ind

icat

esi

gnifi

cant

diff

eren

ces

atP

=0.

05

bet

wee

nth

etw

olo

cati

on

s

OT

U:

Op

erat

ion

alT

axon

omic

Un

it

S15

Table S5. Evaluation of the performance of 2nd-order polynomial models used to estimatedegradation of the different biodegradable plastic mulches.

Thermal Time Calendar TimeMulch MAE NRMSE R2 d MAE NRMSE R2 d

KnoxvilleBioAgri 8.0 13.2 0.79 0.94 8.2 13.3 0.79 0.94Naturecycle 7.5 11.2 0.88 0.97 7.5 11.4 0.88 0.97Organix 5.8 9.1 0.93 0.98 7.0 10.4 0.91 0.98PLA/PHA 4.9 7.6 0.95 0.99 4.8 7.9 0.94 0.98

Mount VernonBioAgri 3.8 11.9 0.79 0.94 3.8 12.2 0.79 0.94Naturecycle 4.5 9.7 0.91 0.98 5.1 10.8 0.88 0.97Organix 2.8 10.9 0.88 0.97 2.7 10.9 0.87 0.97PLA/PHA 2.9 13.4 0.76 0.92 3.2 14.0 0.75 0.91

MAE: mean absolute error (MAE); NRMSE: normalized root mean squared error (normalizedto the range of observed data); R2: coefficient of determination; d: index of agreement

S16

0

25

50

75

100

0 20 40 60 80 100

Bio

degr

adat

ion

(%)

Time (d)

PLA/PHA

BioAgri

Paper (Cellulosic)

Fig. S1. Biodegradation as determined by CO2 release from biodegradable plastic mulches(BioAgri and PLA/PHA) and paper mulch (cellulose) during soil incubation under controlledlaboratory conditions (ASTM D-5988).

S17

250

µm

250

µm

1 m

m1

mm

Bio360 Paper PLA

Uncleaned m

eshbags

after composting

Cleaned meshbags

after composting

Unused

mulches

Unused meshbags

Mes

hbag

sive

size

Fig. S2. Visual observation of BioAg 360, paper, and PLA mulches after composting for 18weeks in 250 µm and 1 mm nylon meshbags in pilot study. Note that the PLA mulch forthis particular experiment was white. Round and rectangular tokens served as identificationtags.

S18

0

25

50

75

100

Tem

pera

ture

(°C

)

0

25

50

75

100

Nov Dec Jan Feb Mar

Tem

pera

ture

(°C

)

Date

(b) 2016

(a) 2015

Atmosphere Compost

Fig. S3. Mean daily atmospheric and compost temperature at 60-cm depth during compostingin (a) 2015 and (b) 2016 at Puyallup, WA. Gray ribbon represents the standard deviationof the mean, with n = 8, except in 2015 where one sensor was removed after each meshbagsampling date. Brown dotted horizontal line at 60oC indicates upper limit for growth of mostthermophilic fungi (Trautmann and Krasny, 1998).

0

25

50

75

100

2016−01 2017−01 2018−01 2019−01

Wat

er a

mou

nt (m

m)

PrecipitationIrrigation

0

25

50

75

100

2016−01 2017−01 2018−01 2019−01

Wat

er a

mou

nt (m

m)

PrecipitationIrrigation

−10

0

10

20

30

40

2016−01 2017−01 2018−01 2019−01Date (yyyy−mm)

Tem

pera

ture

(°C

)

−10

0

10

20

30

40

2016−01 2017−01 2018−01 2019−01Date (yyyy−mm)

Tem

pera

ture

(°C

)

(a) Knoxville

c) Knoxville

(b) Mount Vernon

(d) Mount Vernon

Fig. S4. Daily precipitation (rainfall and irrigation) and daily-averaged air temperature atKnoxville (a,c) and Mount Vernon (b,d) during the 36-month field study.

S19

0.0

0.2

0.4

0.6

0.8

2016−01 2017−01 2018−01 2019−01

Wat

er c

onte

nt (m

3 m−3

)

0.0

0.2

0.4

0.6

0.8

2016−01 2017−01 2018−01 2019−01

Wat

er c

onte

nt (m

3 m−3

)

0

10

20

30

40

2016−01 2017−01 2018−01 2019−01Date (yyyy−mm)

Tem

pera

ture

(°C

)

0

10

20

30

40

2016−01 2017−01 2018−01 2019−01Date (yyyy−mm)

Tem

pera

ture

(°C

)

(a) Knoxville

(c) Knoxville

(b) Mount Vernon

(d) Mount Vernon

PaperBioAgri

NaturecycleOrganix

PLA/PHAPolyethylene

Fig. S5. Daily-averaged soil water content and soil temperature at 10-cm depth at Knoxville(a,c) and Mount Vernon (b,d) during the 36-month field study. Discontinued lines representsperiods when the sensors were temporarily removed for field operations (tillage). Horizontalbars in plot (b) indicate time periods when the soil was flooded.

S20

Fig. S6. Microbial taxa distribution (Class level) at Knoxville and Mount Vernon for thedifferent mulch treatments in Fall 2015. Relative abundances above a cut-off level of 2%are indicated. “Bacteria unclassified” denote taxa with relative abundance above the cut-offlevel of 2%, but that could not be classified.

S21

0 5000 10000 150000

20

40

60

80

100

Mul

ch d

egra

datio

n (%

)

0 10 20 30 400

20

40

60

80

100

Mul

ch d

egra

datio

n (%

)

0 5000 10000 150000

20

40

60

80

100

Mul

ch d

egra

datio

n (%

)

0 10 20 30 400

20

40

60

80

100

Mul

ch d

egra

datio

n (%

)

0 5000 10000 150000

20

40

60

80

100

Mul

ch d

egra

datio

n (%

)

0 10 20 30 400

20

40

60

80

100

Mul

ch d

egra

datio

n (%

)

0 5000 10000 150000

20

40

60

80

100

Thermal time (°C-day)

Mul

ch d

egra

datio

n (%

)

0 10 20 30 400

20

40

60

80

100

Time (months)

Mul

ch d

egra

datio

n (%

)

Knoxville Mount Vernon

Knoxville:

Mount Vernon:

⇥ x ⇥ x

⇥ x ⇥ x

Knoxville:

Mount Vernon:

⇥ x ⇥ x

⇥ x

Knoxville:

Mount Vernon:

⇥ x ⇥ x

⇥ x

Knoxville:

Mount Vernon:

⇥ x ⇥ x

⇥ x ⇥ x

Knoxville:

Mount Vernon:

⇥ x ⇥ x

⇥ x ⇥ x

Knoxville:

Mount Vernon:

⇥ x ⇥ x

⇥ x

Knoxville:

Mount Vernon:

⇥ x ⇥ x

⇥ x

Knoxville:

Mount Vernon:

⇥ x ⇥ x

⇥ x

BioAgri

Naturecycle

Organix

PLA/PHA

BioAgri

Organix

PLA/PHA

Naturecycle

Fig. S7. Measured and simulated degradation of biodegradable plastic mulches in soil as afunction of thermal (left column) and calendar time (right column). Lines represent fitted2nd-order polynomial model and symbols represent measured mulch degradation data. Errorbars are standard deviations of the mean (n = 4).