How and why does willow biochar increase a clay soil water retention capacity?

Con en s lis s a ailable a ScienceDi ec Biomass and Bioene gy jou nal homepage: www.else ie .com/loca e/biombioe Resea ch pape How and why does willow biocha inc ease a clay soil wa e e en ion capaci y? Kimmo Rasa a,∗ , Jaakko Heikkinen a , Ma kus Hannula b , Kai A s ila c , Sampo Kulju a , Ja i Hy äluoma a a Na u al Resou ces Ins i u e Finland, Luke, FI-31600, Jokioinen, Finland b BioMediTech Ins i u e and Facul y o Biomedical Sciences and Enginee ing, Tampe e Uni e si y o Technology, FI-33101, Tampe e, Finland c Uni e si y o Jy askyla, Depa men o Physics, P.O. Box 35, FI-40014, Finland ARTICLE INFO Keywo ds: Biocha X- ay omog aphy 3D image analysis Helium ion mic oscopy Soil wa e e en ion Plan a ailable wa e ABSTRACT Addi ion o biocha in o a soil changes i s wa e e en ion p ope ies by modi ying soil ex u al and s uc u al p ope ies. In addi ion, in e nal mic ome e -scale po osi y ha is able o di ec ly s o e eadily plan a ailable wa e affec s soil wa e e en ion p ope ies. This s udy shows how p ecise knowledge o he in e nal mic o- me e -scale po e size dis ibu ion o biocha can deepen he unde s anding o he biocha -wa e in e ac ions in soils. The mic ome e -scale po osi y o willow biocha was quan i a i ely and quali a i ely cha ac e ized using X- ay omog aphy, 3D image analysis and Helium ion mic oscopy. The effec o biocha applica ion on clay soil wa e e en ion was s udied by con en ional wa e e en ion cu e app oach. The esul s indica e ha he in- e nal po es o biocha , wi h sizes o a 50 and 10 μm (equi alen po e diame e ), inc eased soil po osi y and he amoun o eadily plan a ailable wa e . A e biocha addi ion, changes in soil po osi y we e de ec ed a po e size egimes 5–10 and 25 μm, i.e. biocha po e sizes mul iplied by ac o 0.5. The de ec ed po e size dis ibu ion o biocha does no p edic di ec ly (1:1 compa ibili y) he changes obse ed in he soil mois u e cha ac e is ics. I is likely ha biocha chemis y and po e mo phology affec biocha -wa e in e ac ions ia e.g. su ace oughness and con ac angle. In addi ion, biocha induced changes in soil s uc u e and ex u e affec ed soil mois u e cha ac e is ics. Howe e , he app oach p esen ed is an a ac i e pa hway o mo e gene alized un- de s anding on how and why biocha in e nal po osi y affec s soil mois u e cha ac e is ics. 1. In oduc ion While biocha is conside ed as a po en ial measu e o seques e ca bon in o soil [1], i s seconda y effec s on soil p ope ies a e o en con o e sial [2]. Biocha has been shown o inc ease soil wa e holding capaci y (WHC), bu no solid unde s anding exis s on he e - ec s o , e.g., soil ype, biocha quali y o clima e. Addi ion o biocha in o a gi en soil changes soil ex u al and s uc u al p ope ies. These changes modi y soil mois u e cha ac e is ics in a specific manne de- pending on biocha ype and soil p ope ies (indi ec mechanism). Howe e , biocha as a highly po ous ma e ial also di ec ly affec s soil wa e holding capaci y ia i s in e nal po osi y. The aim o his pape is o s udy how his in e nal po osi y and po e size dis ibu ion a e ela ed o he amoun o plan a ailable wa e in a clay soil. A la ge numbe o pape s desc ibe biocha effec on soil hyd aulic p ope ies and wa e holding capaci y [3–6] wi h he e ogeneous de- sc ip ions o biocha physical quali y. Physical cha ac e is ics o biocha is mos commonly s udied by gas adso p ion echniques ac- companied by he B unaue -Emme -Telle (BET) modelling o de- e mine he specific su ace a ea [7] o Ba e -Joyne -Halenda (BJH) modelling o de e mine po e size dis ibu ion [8]. Howe e , po e space analysis based on gas adso p ion measu emen s is limi ed o po es smalle han 300 nm, whe eby i does no ell much abou he po osi y in he size ange ha is impo an o plan wa e up ake (i.e. mic o- me e -scale po es). The inadequacy o gas adso p ion s udies and d awbacks ela ed o po osi y measu emen s using me cu y po osime y ha e been ad- d essed in ea lie s udies [9,10]. NMR C yopo ome y ha e also been used o s udy po ous ma e ials a nanome e leng h-scale, howe e , esolu ion in he uppe end is limi ed o ew mic ome e s [11]. Kinney e al. [4] add essed he need o quan i a i e echniques o cha ac e ize mic ome e -scale po es wi hin he biocha s due o hei expec ed im- po ance on soil WHC. All hese s udies suppo heo y ha la ge in- e nal mic ome e -scale po es o biocha may ha e ema kable di ec h ps://doi.o g/10.1016/j.biombioe.2018.10.004 Recei ed 21 Decembe 2017; Recei ed in e ised o m 28 Sep embe 2018; Accep ed 2 Oc obe 2018 ∗ Co esponding au ho . Na u al Resou ces Ins i u e Finland, Luke, P oduc ion sys ems, Bio efine y and Biop oduc s, Finland. E-mail add ess: kimmo. asa@luke.fi(K. Rasa). Biomass and Bioene gy 119 (2018) 346–353 0961-9534/ © 2018 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/BY/4.0/). T effec on soil mois u e cha ac e is ics (indi ec effec e e s o changes in soil ex u e and o ma ion o soil agg ega es h ough bindings be- ween biocha and soil pa icles, see Re . [12]). Howe e , he effec o biocha mic ome e -scale po osi y on soil mois u e cha ac e is ics has no been quan i a i ely analysed. We an icipa e ha he way o wa d in inc easing unde s anding o biocha -wa e in e ac ions goes h ough undamen al esea ch on in- e nal po e s uc u e o biocha s. La ely, X- ay omog aphy and 3D image analysis me hods ha e been adap ed o biocha esea ch. This app oach is able o isualize he in e nal po e s uc u e o biocha , and o di ec ly de e mine po e cha ac e is ics such as po e size dis ibu ion and po e con inui y. In ecen s udies X- ay omog aphy has been used o cha ac e ize po e s uc u e o biocha s de i ed om a ious aw ma e ials (e.g. Pinus syl es is,Miscan hus, Populus spp. L., co onseed hull, hay) wi h esolu ion anging om 0.74 o 21 μm[13–17]. Hy ä- luoma e al. imaged a ious ypes o biocha s and hyd ocha s wi h ca. 1μm esolu ion and ound ha conside able pa o he biocha olume consis o po es in size ange ele an o hyd ological p ocesses and s o age o plan a ailable wa e [18]. In addi ion, hey ound high a ia ion in po osi y, po e size dis ibu ion and s uc u al aniso opy be ween diffe en biocha s. Also agg ega es o biocha -amended soil ha e been imaged wi h X- ay omog aphy, al hough a his leng h-scale X- ay omog aphy esolu ion canno cap u e he in e nal po osi y o biocha [19]. The lack o p ecise da a on mic ome e -scale po osi y migh be one eason ha he impo ance o biocha in e nal po osi y di ec ly a - ec ing he soil wa e holding capaci y has been o e looked. In his s udy we used X- ay mic o omog aphy, 3D image analysis and Helium ion mic oscopy o s udy s uc u al p ope ies o willow biocha wi h emphasis on mic ome e -scale po es con ibu ing o s o age o plan a ailable wa e . Fu he , we conduc ed an incuba ion expe imen o s udy he effec o his specific biocha on soil mois u e cha ac e is ic and o iden i y po e egimes al e ed by biocha amendmen . The e- la ionship be ween biocha po e sys em and i s specificeffec on soil mois u e cha ac e is ics is an a ac i e pa hway owa ds de elopmen o p ecisely ailo ed biocha s aimed o enhance wa e use efficiency. 2. Ma e ials and me hods 2.1. Expe imen al soil Soil o incuba ion expe imen was aken om Ko kanoja long- e m d ainage expe imen si e [20] loca ed in Jokioinen in sou he n Finland (N 60.82°, E 23.51°). The s udy ma e ial was collec ed om he opmos 10 cm laye o an annually ploughed plo in sp ing 2016. The hea y clay soil had 64.8%, 30.5% and 4.7% o clay, sil and sand, espec i ely (mass ac ions de e mined by pipe e me hod [21]). Soil o ganic ma e con en was 9.2% (loss o igni ion a 550 °C), ca bon con en 2.9% (Leco analysa o ), pH 6.3 and elec ical conduc i i y 63.1 μScm −1 (soil o wa e a io 1:5). 2.2. Biocha The biocha used in he expe imen was py olysed in an indi ec ly hea ed pilo -scale ba ch- ype py olysis acili y using Willow s em wood (Salix sp, wi h ba k) as aw ma e ial. Tempe a u e p ofile o py olysis p ocess consis ed o wo s eps. Fi s , he empe a u e was aised o 280 °C, and hen u he o 320 °C wi h a es o 2.2 and 0.2 K min −1 , espec i ely. The elemen al analysis (CHNSO) o biocha was ca ied ou using FLASH 2000 se ies analyse and he esul s we e used o calcula e O:C and H:C a omic a ios. The biocha was analysed o pH (SF EN 13037, 1:5 cha o wa e a io), elec ical conduc i i y (EC, SF EN 13038, 1:5 cha o wa e a io), nu ien s and hea y me als (SFS-EN 13650 Aqua Regia ex ac ion and ICP-measu emen ), BET su ace a ea (ISO 9277:2010(E) B unaue –Emme –Telle (BET) Su ace A ea) and Ash con en (SFS 3008, loss o igni ion a 550 °C). The BET su ace a ea o biocha p oduced was 6.8 ± 0.43 m 2 g −1 . The elemen al mass ac ions o C, H, N, S and O we e 74.0 ± 0.0%, 4.1 ± 0.1%, 0.4 ± 0.0%, 0.0 ± 0.0% and 15.8 ± 0.3%, espec- i ely. The ash mass ac ion o he biocha was 3.13 ± 0.15%. The co esponding mole a ios o H:C and O:C we e 0.66 and 0.16, e- spec i ely. Biocha had low con en o main nu ien s (1.5 g kg −1 , 3.6 g kg −1 , 9.1 g kg −1 o P, K, Ca espec i ely) and hea y me als Cd, Cu and Pb (0.79 mg kg −1 , 6.8 mg kg −1 ,<3mgkg −1 , co esponding limi s a e 1.5, 600 and 100 mg kg −1 , espec i ely) con en s we e below limi s se o soil amendmen s in Finland. 2.3. X- ay omog aphy The X- ay compu ed mic o omog aphy imaging was conduc ed wi h Zeiss X adia Mic oXCT-400 (Zeiss, Pleasan on, CA, USA) de ice. Sou ce ol age was 40 kV and sou ce cu en was 250 μA. The pixel size was 1.14 μm. A 20× objec i e was used wi h 2 binning. 1600 p ojec ions we e aken in ull 360°. Each p ojec ion was exposed wi h X- ays o 3 s. No fil e s we e used in he imaging p ocess. Zeiss XMRecons uc o so wa e wi h he fil e ed back p ojec ion algo i hm was used in he econs uc ion o he image s acks. Fo quan i a i e s uc u al analysis he o iginal g ey-scale image was de-noised and segmen ed in o solid and oid phases. De ails o hese image p ocessing s eps ha e been desc ibed by Hy äluoma e al. [18] and e iewed in Appendix A. The sub olume used in he u he image analysis consis ed o 1.9·10 8 oxels (sample size ca. 0.5 mm). The po osi y o he sample was calcula ed by di iding he numbe o oid oxels by he o al numbe o oxel in he analysed olume. The po e-size dis ibu ion was de e mined using an app oach based on ma hema ical mo phology [22] by successi ely applying mo phological opening ope a ions (sphe e wi h adius was used as he s uc u ing elemen ) on he po e space [23]. The po e size based on mo phological opening closely ela es o he s a iona y dis ibu ion o we ing and nonwe ing fluids in po e space and he ela ed capilla y p essu e ia he Young-Laplace equa ion = p γcosθ 2 c(1) whe e γis he su ace ension, θ he con ac angle, and he po e adius defined as he adius o he s uc u ing elemen used in he mo pho- logical opening [24]. This me hod is pa icula ly sui able o ou pu - poses whe e he po e-size dis ibu ion de e mined om imaged po e space is linked o measu ed soil mois u e cha ac e is ic cu e. The s uc u al aniso opy o he sample was quan ified by calcula ing he deg ee o aniso opy using he me hod based on g ey-scale g adien enso [25]. 2.4. Helium ion mic oscopy Helium ion mic oscopy (HIM) is a no el de elopmen wi hin he amily o scanning beam mic oscopes. Ins ead o using elec ons o beam pa icles, as in con en ional scanning elec on mic oscopes (SEM), helium ions a e used. Helium ions ha e a ocal dep h ha is 5–10 imes la ge han elec ons in SEM, a beam spo size below 0.5 nm, and e y small in e ac ion olume p oducing seconda y elec ons, e- sul ing in images wi h e y high esolu ion. Ano he impo an ad- an age o HIM is i s capabili y o image insula ing ma e ials wi hou cha ging effec s (no me al coa ing is needed) as he posi i e su ace cha ging om helium ions and seconda y elec ons can be neu alized wi h an elec on flood gun. A Zeiss O ion NanoFab Helium ion mic o- scope was used in his wo k. 2.5. Soil sample p epa a ion Expe imen al soil was ai -d ied in oom empe a u e and sie ed K. Rasa e al. Biomass and Bioene gy 119 (2018) 346–353 347 h ough 6 mm mesh. Fou diffe en kinds o samples we e p epa ed o incuba ion expe imen . Hal o he samples had na u al (NAT) ag- g ega e s uc u e (< 6 mm). The soil o o he hal was homogenized (HOM) using a olle -mill and hen sie ed (2 mm) in o de o des oy soil mac o agg ega es. Samples wi h NAT and HOM s uc u e ecei ed 0% o 5% mass ac ion biocha amendmen (on d y ma e basis). Biocha used in he expe imen was c ushed in o fine powde using blende . Al oge he 14 samples we e p epa ed. The incuba ion expe imen was conduc ed in 190 cm 3 (7.1 cm in diame e ) me al cylinde s. One end o each cylinde was co e ed wi h a hin ab ic o p e en sample o loss om he cylinde . The cylinde s we e filled wi h soil o mix u e o soil and biocha . The samples we e gen ly shaken and p essed using 10 kg weigh . Finally, he excess ma- e ial was emo ed by spa ula. 2.6. Incuba ion Soil samples we e sa u a ed wi h wa e o six weeks. Sa u a ed samples we e mo ed o sandbox (Eijkelkamp) and he ma ic po en ial was g adually adjus ed o −10 kPa. The ea e , he samples we e in- cuba ed unde cons an ma ic po en ial (−10 kPa) and empe a u e (5 °C) condi ions o six weeks. Cool incuba ion empe a u e was used o hinde he mic obial ac i i y in he samples. 2.7. Soil mois u e cha ac e is ic cu e Volume ac ion o soil wa e in ma ic po en ial ange om −0.3 o −10 kPa was de e mined using sandbox (Eijkelkamp), whe eas p es- su e pla e ex ac o (Soilmois u e Equipmen Co p.) equipped wi h 100 kPa, 300 kPa ba and 500 kPa ce amic pla es was used o p oduce lowe ma ic po en ials. In each s ep he soil samples we e le o equilib a e be o e de e mina ion o he wa e con en . The lowes ma- ic po en ials −1500 kPa and −4000 kPa, we e c ea ed osmo ically by placing 1 g soil sample in o desicca o s equipped wi h con aine s wi h sa u a ed solu ions o ammonium oxala e ((NH 4 ) 2 C 2 O 4 ) o sodium chlo ide (NaCl), espec i ely. The ea e he samples we e emo ed om desicca o s and he wa e con en o he samples we e de e mined by weighing he samples be o e and a e d ying hem a 105 °C. Equilib a ion ime a ied be ween 1 day (a −0.3 kPa) and 6 weeks (a −1500 and −4000 kPa) depending on he ma ic po en ial. 2.8. Effec o biocha amendmen on soil po osi y Da a used o de e mine he soil mois u e cha ac e is ic cu e (SMC) was also used o s udy he effec o added biocha on he po e size dis ibu ion (See Eq (2). below). Howe e , ins ead o using olume ac ion o soil wa e , he wa e con en s we e exp essed as a io o wa e olume o d y mass o sample. This a io was used o ensu e ha esul s con ain only he po e olume ela ed o biocha o each po e size in e al. Fi s , he soil wa e po en ials we e con e ed o ap- p oxima e po e sizes using he Young-Laplace equa ion (Eq. (1)). While Young-Laplace equa ion is de i ed o s aigh capilla ies wi h ci cula c oss sec ion and does no di ec ly apply o soils wi h mo e complex po e geome y, i can be used o de e mine so-called equi alen po e diame e (EPD) d = 2 . Then, he wa e con en s o he samples we e in e pola ed on 2.75 μm po e size in e al using linea in e pola ion. This inc emen co esponds o he inc emen in size o s uc u e ele- men used in de e mina ion o po e size dis ibu ion om omog aphy images. Finally, he effec o biocha on soil po osi y was calcula ed as ollow: =− −− −− − WWMWWM Δ ()/()/, pp p p p p p p p[, Δ] Δ 5% Δ 0% (2) whe e Δ is he effec o biocha amendmen on specific po e olume o soil in each po e size in e al (cm 3 /g), W is he olume ac ion o soil wa e o he sample (=po e olume, cm 3 ), M is he d y ma e weigh o he sample (g), p is he uppe limi and p-Δp lowe limi o po e size in e al (p = 2.75 μm, 5.5 μm, …;Δp = 2.75 μm) and ˂˃ 5% and ˂˃ 0% ep esen s he a e age o he samples wi h and wi hou biocha . Calcula ion was pe o med sepa a ely o wo diffe en soil s uc u es (HOM, NAT). 2.9. S a is ical analysis The dependence o plan a ailable wa e on he biocha addi ion and soil s uc u e was s udied using wo-way analysis o a iance. Dependen a iables we e p opo ions ( alues be ween 0 and 1) and he e o e logi ans o ma ion was used o make he model assump ion applicable. Fo he p esen a ion o he esul s, he es ima ed means we e ans o med back o he o iginal scale. Bon e oni co ec ion was applied o mul iple compa isons. All s a is ical analyses we e pe o med using Ma lab wi h S a is ics Toolbox. 3. Resul s and discussion 3.1. X- ay omog aphy o willow biocha The isual obse a ion o he X- ay omog aphy image e ealed ha he po e s uc u e o willow biocha py olysed a 320 °C e ained he ini ial s uc u al cha ac e is ics o esh s em wood (Fig. 1). Ob iously, he ascula issues (xylem) in sapwood and/o hea wood appea as biocha po osi y. Analogously o li ing wood issue, hese po es a e Fig. 1. X- ay omog aphic econs uc ion o willow biocha (le ). The 3D isualiza ion o omog aphy da a was p oduced wi h Pa a iew [29] and Blende . The omo- g aphy da a consis ing o he imaging slices we e fi s o a ed wi h ImageJ [30] in a way ha po es a e pe pendicula o he cu ing ace. A e his a cuboid piece was cu om a sample. A 57 μm hick slice cu om he middle o a cuboid piece is shown on igh . K. Rasa e al. Biomass and Bioene gy 119 (2018) 346–353 348 mainly esponsible o wa e s o age and anspo wi hin he biocha . The biocha po e sys em was quan ified by image analysis o he X- ay omog aphy image. The esul s show ha he o al po osi y o biocha is 0.6 and he po e size dis ibu ion is bimodal wi h local maxima a ound diame e s 10 and 50 μm(Fig. 2 op panel). The deg ee o aniso opy was 32.0 (see Hy äluoma e al. [18] o e e ence alues o deg ee o aniso opy o se e al wood based biocha s) which p o es ha po osi y o he biocha is highly aniso opic and consis s o pa allel cylind ical po es wi h minimum numbe o la e al connec ions ( e ified by HIM, see below). In gene al, hese findings a e well in line wi h ea lie published 2D scanning elec on mic oscopy images indica ing bimodal cha ac e is ics o esh willow wood s uc u e e.g. Re s. [26,27]. In addi ion, a ecen pape [28] discussing he effec o py - olysis empe a u e on mic ome e -scale po osi y o willow (Salix schwe inii ‘Amgunskaja') biocha shows simila bimodal po e cha - ac e is ics as in he p esen pape . The used imaging esolu ion p e en s obse a ion o po es in sub- mic ome e size ange. Howe e , he low BET su ace a ea (6.8 ± 0.43 m 2 g −1 ) sugges s poo ly de eloped nanopo osi y o he s udied biocha . This sugges ion is in line wi h p e ious s udies e- po ing de elopmen o py ogene ic nanopo es only in highe py olysis empe a u es [10,31,32]. To e i y his hypo hesis, we s udied nano- po osi y o biocha po e walls using HIM, which indica ed ha cell wall s uc u es emaining in he biocha do no con ain isible nano-scale po es (Fig. 3). HIM obse a ions and deg ee o aniso opy oge he sugges ha wa e s o age and flow wi hin willow biocha akes place in “bundle o cylind ical capilla ies”as is assumed in Eq. (1) (see also discussion in Sec. 3.3). 3.2. Effec s o biocha amendmen on soil-mois u e cha ac e is ics Addi ion o biocha educed he bulk densi y o NAT ( om 0.91 o 0.87 g cm −3 , P = 0.001) and HOM ( om 1.08 o 0.93 g cm −3 , P < 0.001) samples compa ed o samples wi hou biocha . The effec was mo e p onounced in HOM samples, because in HOM samples bio- cha eplaces mo e mine al soil han in he case o NAT. In NAT samples biocha alls pa ially in po es be ween agg ega es and hus affec s he soil bulk densi y less. These indi ec effec s o biocha amendmen ex- plain changes obse ed in he we -end o SMC ( ela ed o po es wi h la ge size han he size o in e nal po es o biocha , see below). Biocha applica ion influenced soil mois u e cha ac e is ics o e he en i e ange o measu ed ma ic po en ials. The majo changes occu ed a ma ic po en ials abo e −316 kPa, while diffe ences le elled ou a he d y-end o he cu e (Fig. 4). Acco ding o Young-Laplace equa ion, in ma ic po en ial o −300 kPa wa e emaining in soil is s o ed in po es wi h EPD < 1 μm. Biocha addi ion hus modifies soil po osi y especially a po e size egime EPD > 1 μm. In gene al, his finding in- dica es ha biocha wi h high mic ome e -scale po osi y, obse ed wi h X- ay omog aphy and 3D image analysis, eflec s di ec ly o soil WHC in egime impo an o plan wa e up ake (Fig. 4,Table 1). Close inspec ion o SMC o NAT and HOM samples (Fig. 2) e eals ha biocha applica ion inc eased po osi y in wo specific egimes. In bo h soils (i espec i e o soil s uc u e), a clea inc ease in po osi y occu ed a po e sizes a ound 25 μm (EPD). The o he maximum is app oxima ely a EPD 5 μm and 10 μm o NAT and HOM samples, e- spec i ely. These peaks co espond he po e sizes obse ed in he bio- cha (10 and 50 μm, EPD) mul iplied by ac o 0.5 o la ge po es. This ac o diffe s sligh ly be ween NAT and HOM samples o smalle po e class. The ac ha a bimodal shape is obse ed bo h in he po e-size dis ibu ion de e mined wi h image analysis and de i ed om SMC p o ide s ong suppo o he in e ence ha biocha affec s he soil mois u e cha ac e is ics g ea ly ia di ec mechanism. In nex sec ion some biocha - ela ed p ope ies affec ing how biocha in e nal po osi y ansla es o changes obse ed in SMC a e discussed in mo e de ails. 3.3. Biocha po e sizes s. changes in soil mois u e cha ac e is ics The e a e se e al easons why i is unlikely ha a one- o-one compa ibili y be ween 3D image analysis and SMC esul s will be ob- ained. Con e sion o he SMC o po osi y esponse cu e was by using he Young-Laplace equa ion wi h gene alized assump ions which a e ypically no ully alid o any p ac ical po ous ma e ial including he po e sys em conside ed he e. One assump ion is ha po e space consis s o a bundle o cylind ical capilla ies wi h uni o m ci cula c oss sec ion o diffe en sizes and pe ec ly smoo h su aces. In addi ion i is as- sumed ha po e walls a e chemically homogeneous and pe ec ly we ing, i.e., in he Young-Laplace equa ion cos θ= 1 a all po e su - aces. Fig. 2. Po e size dis ibu ion de e mined by 3D image analysis o X- ay omo- g aphy image ( op panel) and he change in he po e size dis ibu ion due o biocha addi ion o NAT (middle panel) and HOM (bo om panel) samples. Effec o biocha on soil po osi y (Δ , cm 3 g −1 soil) was de e mined by sub- ac ing he soil wa e con en (exp essed as a a io o wa e olume o d y mass o sample) o pu e soil om ha o soil-biocha mix u e as desc ibed in Sec. 2.8. Ma ic po en ial was con e ed o equi alen po e size using Eq. (1). K. Rasa e al. Biomass and Bioene gy 119 (2018) 346–353 349 P ac ically all po e su aces a e ough. Wenzel's model is widely used o desc ibe he effec o su ace oughness on he con ac angle [33]. Ano he effec o su ace he e ogenei ies is so-called con ac angle hys e esis which means ha con ac angle is no unique bu a ies be ween wo ex eme alues known as eceding and ad ancing con ac angles. Also, assump ion o pe ec ly we ing po e walls is no ealis ic. To accoun o bo h non-pe ec we ing and oughness, one can w i e = ∗ cos θ cos θ (3) whe e is he oughness ac o defined as he a io o he ue su ace a ea o he appa en (p ojec ed) a ea and θ* is so-called in insic con- ac angle. In eali y biocha s a e subc i ically hyd ophobic (0° < θ* < 90°) and po e su aces ough, whe eby cos θ* < 1 and > 1. The e o e, depending on he physical and chemical p ope ies o po e walls, po e sizes can be ei he la ge o smalle han hose deduced om SMC. The impo ance o we ing p ope ies on he hyd ologic effec s o biocha on soil has been demons a ed by Suliman e al. [34]. I mus also be emphasized ha simple app oaches o we abili y (including he discussion abo e) do no conside complica ed ime-de- penden beha io obse ed in many na u al sys ems. We abili y o many na u al ma e ials, including soil and biocha , is a highly complex phenomenon which is no ye p ope ly unde s ood. In p ac ice, i is difficul o de e mine he con ac angle needed o p ope ly use he Young-Laplace equa ion. The commonly used me hod ha is based on op ical de e mina ion o he con ac angle o a sessile d op does no measu e in insic con ac angle as he measu emen is pe o med on a ough su ace o ma e ial. Sessile d op me hod has been used o de e mine con ac angle o biocha s [15,35]. I is impo an o no e ha his la ge-scale oughness does no co espond o he small- scale oughness used in Eq. (2). The in insic con ac angle should be de e mined a he h ee-phase con ac line wi hin he po e which is no possible in p ac ice. Fu he mo e, he de i ed oughness ac o in Eq. (2) is he oughness a po e walls which also is p ac ically impossible o de e mine. The discussion abo e explains he ac o s why he e is no one- o- Fig. 3. Helium ion mic oscopy images o willow biocha . Size o he images (a) 1 mm, (b) 50 μm, (c) 4μm, and (d) 2 μm. In image (a) o e all po osi y o biocha is shown while image (b) is ocused on po es wi h EPD o app ox. 5–10 μm. Image (c) shows po e wall be ween he biocha po es and (d) is he same po e wall wi h highe magnifica ion indica ing ha he e is no isible connec ions be ween indi idual po es. Fig. 4. Soil mois u e cha ac e is ics cu e o each s udied soil s uc u e and biocha combina ions. P esen ed esul s a e a e age o each ea men s (n = 4 and n =3 o samples wi h and wi hou biocha addi ion, espec i ely). E o ba s deno e s anda d de ia ions. K. Rasa e al. Biomass and Bioene gy 119 (2018) 346–353 350 one co espondence be ween he po e sizes de e mined by image ana- lysis and hose de i ed om SMC. We also obse ed ha he e is di - e en ans o m coefficien be ween he la ge po es and small po es. One explana ion could be he exis ence o cons ic ions o h oa s in he biocha po es ha could be o med, e.g., by a compounds which could pa ially clog he po es. Image analysis and po osi y esponse de i ed om SMC can be di- ec ly compa ed only i all po es a e connec ed o he sample su ace di ec ly o h ough la ge po es. The Image analysis de e mines po e size dis ibu ion while SMC me hod is ela ed o he size dis ibu ion o po e h oa s. E en hough he po es in he used biocha we e ubula and did no ha e ob ious h oa s (see Secs. 3.1 and 3.2), we ne e - heless es ima ed he effec o po e h oa s on he compa ison by de- e mining he h oa size o po e size a io using he me hod desc ibed in Appendix B. In he analysis, we sepa a ed he wo po e size classes p esen in he s udied biocha by using po e diame e 25 μm as h eshold be ween la ge and smalle po es. The a e age h oa size o po e size a ios we e 0.97 and 0.78 o he la ge and smalle po es, espec i ely. Thus o la ge po es he h oa effec was p ac ically negligible while o smalle po es a clea effec was obse ed. This explains a leas pa ly why he posi ions o modes ela ed o smalle and la ge po es a e no shi ed by same ac o when compa ing SMC esul s o po e size dis- ibu ion de e mined by image analysis (Fig. 2). Al hough he hypo he ical conside a ion o biocha di ec effec on he SMC is well es ablished, he me hodological limi a ions ha e hin- de ed s udies on his issue. Howe e , he SMC esul s o sandy soils and SEM images o biocha p esen ed by Abel e al. [36] could possibly be in e p e ed in simila way han in ou s udy. Liu e al. ound ha bio- cha amendmen shi ed soil po e size dis ibu ion owa d smalle po es, bu biocha effec was no unambiguously es ablished [5]. In addi ion o he di ec influence o biocha in e nal po e sys em, biocha affec s soil mois u e cha ac e is ics ia indi ec mechanisms, i.e. soil-biocha in e ac ions affec ing soil s uc u e (agg ega e o ma- ion) and/o changes in soil pa icle size dis ibu ion [6,37]. The effec o soil s uc u al changes p obably explains why he modes o HOM ea men a e wide han hose o NAT ea men (Fig. 2). In e ac ions be ween soil-soil and soil-biocha a e mo e common in HOM ea men wi h des oyed soil agg ega es han in NAT ea men whe e biocha ac s mo e independen ly in ca i ies be ween exis ing soil agg ega es. In any case, obse ed bimodal po e size dis ibu ion o willow biocha and consequen inc ease in soil wa e e en ion p ope ies a co esponding po e size egimes (mul iplied by ac o 0.5) p o ides s ong e idence ha biocha in e nal mic ome e -scale po osi y con- ibu es di ec ly o soil wa e s o age. This ou come encou ages u he s udies and me hodological de elopmen in o de o imp o e unde - s anding on ela ionship be ween biocha mic on scale po osi y and i s di ec effec on soil mois u e cha ac e is ics. 3.4. Effec s o biocha amendmen on soil plan a ailable wa e Biocha applica ion inc eased he plan a ailable wa e (PAW) by 17% and 32% o HOM and NAT samples in compa ison o samples wi h no biocha addi ion (Table 1). I is no able ha in bo h s udied soil s uc u es he biocha induced inc ease in PAW occu ed mos ly ma ic po en ial egime om −10 o −316 kPa (Fig. 2), conside ed he e as eadily plan a ailable wa e (RPAW), which co esponds o po es wi h EPD > 1 μm. Wi h espec o c op p oduc ion a d y condi ions and/o du ing igo ous g ow h, inc eased s o age o RPAW helps plan s o main ain po en ial anspi a ion a e o longe ime. I is well documen ed ha d y ma e p oduc ion o plan s i ually ceases when hey lose u go p essu e and his may occu a be o e pe manen wil ing poin (PWP) [38]. Meye and G een epo ed ha whea could use 52–57% o PAW be o e educ ion in g ow h became e iden [39]. In wide pe spec i e, measu es like biocha amendmen o inc ease wa e p oduc i i y in ag icul u e (i.e. ou pu yield pe uni wa e used) a e o impo ance in he s uggle agains wa e sca ci y h ea ening global ood secu i y [40]. The p esen s udy highligh s he impo ance o a di ec link be ween he biocha mic ome e -scale po osi y and RPAW. I espec i ely o soil s uc u e, biocha amendmen inc eased soil po osi y a he same po e size egimes (EPD 5–10 and 25 μm, Fig. 2), which e e o biocha in- duced changes ia di ec mechanism (biocha in e nal po osi y). I could be hypo hesized ha his specific biocha would esul in simila changes in o he soil ypes as well. I he hypo hesis holds, he app oach p esen ed abo e could p o ide a ascina ing ool-box enabling de el- opmen o ailo ed biocha s ha could modi y soil mois u e cha - ac e is ics a well-defined mois u e egimes o e he ange o soil ypes. Such ailo ed biocha p oduc s could imp o e biocha pe o mance in c op p oduc ion a diffe en clima ic a eas, soil ypes and p oduc ion sys ems, which is essen ial when sus ainabili y o biocha is conside ed in global scale [1]. In any case, his hypo hesis calls o u he ca e ul s udies. Howe e , i should also kep in mind ha while biocha effec s on SMC ia di ec mechanism may be independen o soil ype, he biocha induced changes in soil bulk densi y, ex u e and s uc u e a e s ongly depend on he p ope ies o ecei ing soil and he ype o biocha used. 4. Conclusions The p esen s udy add esses he impo ance o biocha mic ome e - scale po osi y when used as soil amendmen . I was shown ha biocha in e nal po osi y de ec ed by X- ay omog aphy modifies SMC ia di ec mechanism. Bimodal po osi y o willow biocha ac ed as aceable ma ke when added o a clay soil. Howe e , a co ec ion ac o o p edic ela ionship be ween biocha po e size dis ibu ion and i s ob- se ed effec s on SMC was needed (in his case close o 0.5) due o e.g. oughness and hyd ophobici y o biocha po es. Because biocha was shown o inc ease soil po osi y due o i s in e nal po osi y (di ec me- chanism) i could be hypo hesized ha his effec is independen om soil ype. On he o he hand, biocha induced changes in soil ex u e and s uc u e, and modified SMC simul aneously (indi ec mechanism). The la e mechanism is known o be s ongly dependen on soil and biocha p ope ies, which may explain why esul s on biocha effec on SMC a e o en con o e sial. F om he p ac ical poin o iew, he willow biocha used in his expe imen can g ea ly inc ease amoun o eadily plan a ailable wa e (be ween ma ic po en ials −10 and −316 kPa) in clay soil i espec i e o soil ini ial s uc u e. The esul sugges s ha biocha amendmen Table 1 Wa e olume ac ion a ma ic po en ial o −10 kPa (Field capaci y, FC) and −1500 kPa (Pe manen wil ing poin , PWP), plan a ailable wa e (PAW) be ween hese ma ic po en ials and eadily plan a ailable wa e (RPAW) be ween ma ic po en ials o −10Kpa and −316 kPa. FC (cm 3 cm −3 ) PWP (cm 3 cm −3 ) PAW (cm 3 cm −3 ) RPAW (cm 3 cm −3 ) NAT 0% 0.36 p < 0.001 0.13 p = 0.01 0.22 p < 0.001 0.08 p = 0.02 5% 0.41 0.12 0.29 0.14 HOM 0% 0.45 p = 0.04 0.15 p < 0.001 0.30 p < 0.001 0.07 p < 0.001 5% 0.48 0.12 0.35 0.18 K. Rasa e al. Biomass and Bioene gy 119 (2018) 346–353 351 would help plan s o wi hs and mois u e s ess du ing d y pe iods and/ o igo ous g ow h s a e. Wi h espec o biocha use in g owing media o ag onomic applica ions, we s ongly ecommend u iliza ion o e- sea ch me hods capable o e eal mic ome e -scale po e s uc u e o biocha s. Acknowledgemen s Riikka Keskinen, Johanna Nikama and Ilkka Sa ikka a e wa mly hanked o ca ying ou he labo a o y wo k ela ed o he soil mois u e cha ac e is ic cu e. We also wish o hank Mi ja Muhola and A e Mikkelson om VTT Technical Resea ch Cen e o Finland L d o ca ying ou BET su ace a ea and CHNSO analysis. This p ojec has ecei ed unding om he Eu opean Union's Ho izon 2020 esea ch and inno a ion p og amme unde g an ag eemen No 637020 −MOBILE FLIP. Appendix A Image p ocessing G ey-scale X- ay omog aphy image was fil e ed wi h a h ee-dimensional median fil e ( adius 2). Fil e ed image was segmen ed in o po e and solid oxels by global h esholding. Th eshold alue was selec ed wi h he modified O su's me hod [41] desc ibed Hy äluoma e al. [18]. This modified e sion u ilizes ideas p esen ed ea lie by Hapca e al. [42]. Segmen ed image was hen fil e ed wi h a majo i y fil e ( adius 2) and finally isola ed (“floa ing”) objec s wi h olume less han 1000 oxels we e emo ed om he image. A c oss sec ion o he g ey-scale image and co e- sponding c oss-sec ion o he segmen ed and fil e ed image is shown in Fig. A1. Figu e A1. C oss sec ions o he x- ay omog aphy images o he willow biocha . On le , g ey-scale image wi h g ey-scale his og am in he inse . Ligh e colo in he his og am shows he g ey-scale alues in e p e ed as po e space and da ke hose in e p e ed as solid ma e ial. On igh , final segmen ed and fil e ed image used in analyses. Appendix B Po e h oa analysis The a io o po e size and he smalles h oa leading o ha po e was es ima ed in he ollowing way. As a s a ing poin , a omog aphic image ha was segmen ed in o po es and solids was used. Me hod u ilizes wo scala fields and one ec o field de i ed om he image. Fi s is dis ance ans o m which de e mines he sho es dis ance o he po e walls o each po e oxels. Dis ance ans o m was compu ed using so-called d [3–5,7] Cham e me ics wi h a escaling p ocedu e as explained in de ail by S ensson and Bo ge o s [43]. Secondly, o each oxel a po e size alue was de e mined by mo phological opening, e.g. Re . [24]. Finally, we compu ed po e-scale flow field o sa u a ed p essu e-d i en flow wi h d i ing p essu e pa allel wi h he po e di ec ion. He e he la ice Bol zmann me hod was u ilized o flow compu a ion [44]. D3Q19 la ice Bol zmann model wi h wo- elaxa ion ime collision ope a o was used and he no-slip bounda y condi ion a fluid-solid bounda ies was en o ced wi h bounce-back bounda y condi ion. Sample size in he simula ion was 454 × 400 × 638 μm 3 . Fu he de ails o he used me hod can be ound om, e.g., [45]. Th oa size was hen de e mined by u ilizing dis ance ans o m, po e size and flow fields. Inspec ion poin s we e selec ed as local maxima o dis ance ans o m which we e de e mined using a 5 × 5 × 5 mo ing window. Fo hese poin , he s eamlines we e in eg a ed bo h ups eam and downs eam, and he smalles po e size ound ollowing he s eamline was hen de- e mined in bo h di ec ions. As poin s along s eamline do no coincide wi h he g id nodes, ilinea in e pola ion was u ilized o de e mine he po e-size alues. This me hod p oduces wo h oa sizes, one ups eam and ano he downs eam om he inspec ion poin . La ge o hese h oa s was selec ed as i is he easie ou e o he po e o be d ained. The h oa size hus de e mined is compa ed o he po e size a he inspec ion poin o de e mine he h oa o po e size a io. I should be no ed ha , as he omog aphy image used in his analysis is smalle han he g ain size used in he expe imen , he esul s ob ained a e only app oxima e. In eal biocha g ains ou es o su ace a e longe and hus smalle h oa s could be p esen . The e o e, he esul s ob ained can sligh ly unde es ima e he effec o po e h oa s. K. Rasa e al. Biomass and Bioene gy 119 (2018) 346–353 352 Re e ences [1] D. Wool , J.E. Amone e, F.A. S ee -Pe o , J. Lehmann, S. Joseph, Sus ainable biocha o mi iga e global clima e change, Na . Commun. 1 (56) (2010) 1–9 Na . Commun. [2] A. Mukhe jee, R. Lal, Biocha impac s on soil physical p ope ies and g eenhouse gas emissions, Ag onomy 3 (2014) 313–339. [3] R.T. Ba nes, M.E. Gallaghe , C.A. Masiello, Z. Liu, B. Dugan, Biocha -Induced changes in soil hyd aulic conduc i i y and dissol ed nu ien fluxes cons ained by labo a o y expe imen s, PLoS One 9 (9) (2014) e108340, , h ps://doi.o g/10. 1371/jou nal.pone.0108340. [4] T.J. Kinney, C.A. Masiello, B. Dugan, W.C. Hockaday, M.R. Dean, K. Zygou akis, R.T. Ba nes, Hyd ological p ope ies o biocha s p oduced a diffe en empe a- u es, Biomass Bioene 49 (2012) 34–43. [5] C. Liu, H. Wang, X. Tang, Z. Guan, B.J. Reid, A.U. Rajapaksha, Y.I. Ok, H. Sun, Biocha inc eased wa e holding capaci y bu accele a ed o ganic ca bon leaching om a sloping a mland soil in China, En i on. Sci. Pollu . R. 23 (2016) 995–1006. [6] Z. Liu, B. Dugan, C.A. Masiello, H.M. Gonne mann, Biocha pa icle size, shape, and po osi y ac oge he o influence soil wa e p ope ies, PLoS One 12 (6) (2017) e0179079h ps://doi.o g/10.1371/jou nal.pone.0179079. [7] S. B unaue , P. Emme , E. Telle , Adso p ion o gases in mul imolecula laye s, J. Am. Chem. Soc. 60 (1938) 309–319. [8] E. Ba e , L. Joyne , P. Halenda, The de e mina ion o po e olume and a ea dis- ibu ions in po ous subs ances. I. Compu a ions om ni ogen iso he ms, J. Am. Chem. Soc. 73 (1951) 373–380. [9] M. G ay, M.G. Johnson, M.I. D agila, M. Klebe , Wa e up ake in biocha s: he oles o po osi y and hyd ophobici y, Biomass Bioene 61 (2014) 196–205. [10] C. B ewe , V. Chuang, C. Masiello, H. Gonne mann, X. Gao, B. Dugan, L. D i e , P. Panzacchi, K. Zygou akis, C. Da ies, New app oaches o measu ing biocha densi y and po osi y, Biomass Bioene g. 66 (2014) 176–185. [11] J.B.W. Webbe , P. Co be , K.T. Semple, U. Ogbonnaya, W.S. Teel, C.A. Masiello, Q.J. Fishe , J.J. Valenza II, Y.-Q. Song, Q. Hu, An NMR s udy o po ous ock and biocha con aining o ganic ma e ial, Mic opo . Mesopo . Ma . 178 (2013) 94–98. [12] F.G.A. Ve heijen, S. Jeffe y, A.C. Bas os, M. an de Velde, I. Dia as, Biocha Applica ion o Soils. A C i ical Scien ific Re iew o Effec s on Soil P ope ies, P ocesses and Func ions. EUR 24099 EN, Office o he Official Publica ions o he Eu opean Communi ies, Luxembou g, 2010, p. 149. [13] M.I. Bi d, P.L. Ascough, I.M. Young, C.V. Wood, A.C. Sco , X- ay mic o omog aphic imaging o cha coal, J. A chaeol. Sci. 35 (2008) 2698–2706. [14] P. Con e, N. Nes le, Wa e dynamics in diffe en biocha ac ions, Magn. Reson. Chem. 53 (2015) 726–734. [15] S. Jeffe y, M.B.J. Meinde s, C.R. S oo , T.M. Bezeme , T.F.J. an de Voo de, L. Momme , J.W. an G oenigen, Biocha applica ion does no imp o e he soil hyd ological unc ion o a sandy soil, Geode ma 251–252 (2015) 47–54. [16] K. Jones, G. Ramak ishnan, M. Uchimiya, A. O lo , New applica ions o X- ay o- mog aphy in py olysis o biomass: biocha imaging, Ene g. Fuel. 29 (2015) 1628–1634. [17] L.S. Schnee, S. Knau h, S. Hapca, W. O en, T. Eickho s , Analysis o physical po e space cha ac e is ics o wo py oly ic biocha s and po en ial as mic ohabi a , Plan Soil 408 (2016) 357–368. [18] J. Hy äluoma, S. Kulju, M. Hannula, H. Wikbe g, A. Källi, K. Rasa, Quan i a i e cha ac e iza ion o po e s uc u e o se e al biocha s wi h 3D imaging, En i on. Sci. Pollu . Res. 25 (2018) 25648–25658. [19] X. Yu, C. Wu, Y. Fu, P.C. B ooke s, S. Lu, Th ee-dimensional po e s uc u e and ca bon dis ibu ion o mac oagg ega es in biocha -amended soil, Eu . J. Soil Sci. 67 (2016) 109–120. [20] E. Tu ola, A. Paajanen, Influence o imp o ed subsu ace d ainage on phospho us losses and ni ogen leaching om a hea y clay soil, Ag ic. Wa e Manag. 28 (1995) 295–310. [21] P. Elonen, Pa icle-size analysis o soil, Ac a Ag al. Fenn. 122 (1971). [22] G. Ho gan, Ma hema ical mo phology o analysing soil s uc u e om images, Eu . J. Soil Sci. 49 (1998) 161–173. [23] M. Hilpe , R. Glan z, C.T. Mille , Calib a ion o a po e-ne wo k model by a po e- mo phological analysis, T anspo Po ous Med. 51 (2003) 267–285. [24] M. Hilpe , C.T. Mille , Po e-mo phology-based simula ion o d ainage in o ally we ing po ous media, Ad . Wa e Resou . 24 (2001) 243–255. [25] Z. Tabo , E. Roki a, Quan i ying aniso opy o abecula bone om g ay-le el images, Bone 40 (2007) 966–972. [26] J. F omm, Ho monal physiology o wood g ow h in willow (Salix iminalis L.): e - ec s o spe mine and abscisic acid, Wood Sci. Technol. 31 (1997) 119–130. [27] W. Schoch, J. Helle , F.H. Schweing ube , F. Kienas , Wood Ana omy o Cen al Eu opean Species, (2004) Online e sion: www.woodana omy.ch. [28] J. Hy äluoma, M. Hannula, K. A s ila, H. Wang, S. Kulju, K. Rasa, Effec s o py - olysis empe a u e on he hyd ologically ele an po osi y o willow biocha , J. Anal. Appl. Py ol. h ps://doi.o g/10.1016/j.jaap.2018.07.011. [29] J. Ah ens, B. Ge eci, C. Law, Pa aView: an End-use Tool o La ge Da a Visualiza ion, Visualiza ion Handbook, Else ie , 2005 ISBN-13: 978-0123875822. [30] C.A. Schneide , W.S. Rasband, K.W. Elicei i, NIH Image o ImageJ: 25 yea s o image analysis, Na . Me hods 9 (2012) 671–675. [31] J. Pa k, I. Hung, Z. Gan, O. Rojas, K. Lim, S. Pa k, Ac i a ed ca bon om biocha : influence o i s physicochemical p ope ies on he so p ion cha ac e is ics o phe- nan h ene, Bio esou . Technol. 149 (2013) 383–389. [32] J. Zhang, C. You, Wa e holding capaci y and abso p ion p ope ies o wood cha s, Ene g. Fuel. 27 (2013) 2643–2648. [33] R.N. Wenzel, Resis ance o solid su aces o we ing by wa e , Ind. Eng. Chem. 28 (1936) 988–994. [34] W. Suliman, J. Ha sh, N. Abu-Lail, A. Fo una, I. Dallmeye , M. Ga cia-Pé ez, The ole o biocha po osi y and su ace unc ionali y in augmen ing hyd ologic p op- e ies o a sandy soil, Sci. To al En i on. 574 (2017) 139–147. [35] G. Ojeda, S. Ma ana, A. À ila, J. Alcañiz, M. Volkmann, J. Bachmann, A e soil- wa e unc ions affec ed by biocha applica ion? Geode ma 249–250 (2015) 1–11. [36] S. Abel, A. Pe e s, S. T inks, H. Schonsky, M. Facklam, G. Wessolek, Impac o biocha and hyd ocha addi ion on wa e e en ion and wa e epellency o sandy soil, Geode ma 202–203 (2013) 183–191. [37] T.J. Lim, K.A. Spokas, G. Feye eisen, J.M. No ak, P edic ing impac o biocha addi ions on soil hyd aulic conduc i i y, Chemosphe e 142 (2016) 136–144. [38] O.T. Denmead, R.H. Shaw, A ailabili y o soil wa e o plan s as affec ed by mois u e con en and me eo ological condi ions, Ag on. J. 54 (1962) 385–390. [39] W.S. Meye , G.C. G een, Wa e use by whea and plan indica o s o a ailable wa e , Ag on. J. 72 (1980) 253–257. [40] F.R. Rijsbe man, Wa e sca ci y: ac o fic ion? Ag ic. Wa e Manag. 80 (2006) 5–22. [41] N. O su, A h eshold selec ion me hod om g ay-le el his og ams, IEEE T. Sys . Man. Cyb. 9 (1979) 62–66. [42] S.M. Hapca, A.N. Hous on, W. O en, P.C. Ba eye, New local h esholding me hod o soil images by minimizing g ayscale in a-class a iance, Vadose Zone J. 12 (2013), h ps://doi.o g/10.2136/ zj2012.0172. [43] S. S ensson, G. Bo ge o s, Digi al dis ance ans o ms in 3D images using in- o ma ion om neighbou hoods up o 5x5x5, Compu . Vis. Image Und. 88 (2002) 24–53. [44] S. Succi, The La ice Bol zmann Equa ion o Fluid Dynamics and beyond, Cla endon P ess, Ox o d, 2001, p. 288. [45] J. Hy äluoma, V. Niemi, M. Thapaliya, E. Tu ola, J. Jä ns ed , J. Timonen, La ice Bol zmann simula ion o flow-induced wall shea s ess in po ous media, T ansp. Po . Med. 121 (2018) 353–368. K. Rasa e al. Biomass and Bioene gy 119 (2018) 346–353 353