• Land cover change analysis : The relevant third-party deforestation risk analysis pq ur oevs it di oenr : c“oWn dh ui ccht sf aarnmasn ha al yvsei ss et oe na nnsown e- cr osme vpel ri aa nl tk ed ye fqour ee ss tt ai ot inosn, i anfct leur d2i n0 g2 0t h? ”e. T h e me xeptehrot droe lvoi ge yw ctoomr ebdi nuecse vt ehreyr hi si kg ho-fr feaslos leupt ioosni tsi va et es l. lTi the ei mp raogveirdye, rg rr oe ul i ne sd -otnr ustaht ienl lgi ,t ea n d monitoring and land-cover change models, including: - 2020 Base Land Cover Model : This deep learning model uses diverse imagery sf ooruerscte, sc oa ct odai ,f fceorfef en et , rpe as ol ml u ot iiol ,nrsutbobcelra, sasni fdy wl aant de rt. yTphees mi noldoec la il sc do enst iegxnt se ,ds tuoc hp earsf o r m rreellaiatbedlytaoccrloosusdvacorivoeursosremnsiossrisnagnddarteas. olutions, and to overcome challenges - Annual Change Detection Model : This takes as input the land cover outputs before awnhde tahf teerr daegf oi vreens tyaet iaorn, aol oc cnugrwr eidt hdau rs ienrgi etshoef yiemaar goefsi ni nt ebreetswt . e e n . I t t h e n e s t i m a t e s Tn oa gt ue rt eh eorf, at hneysseumb soedqeul se nwt oargkr itcouildt ue rnat li flya nd de f uo sr ee s(tcaot ci oona af acrt mi v ii tnyg a, on rd aens ot at hb el i rs hc rtohpe) . • Manual review : The initial modelled results are subject to rigorous manual review by torrai gi ni ne ds . aTnhaelsyes trse wv i ietwh sr ecloenv `ai nr mt r ewghi oe nt hael re xdpe ef or rt iesset ai nt i oc no f of ecec,ucror ceodaa, na dn dv aplai dl ma t-ep trho ed u c i n g sa un ba sl yesqi su epnr ot vl ai dn edrutsoe i. dTehni tsi fpyr loocceastsi oanl lso ww hs et hr ee froerl ee svta wn tatshsi rpde-cpi `ai rc tayl l yd ecfoonr ve es trat et ido nt or i s k Ev eUrDi ` Ri e- sr et shtarti ct theed pcoolmy gmo no sd idt oi e ns ,ostui cnht earss ec cotf foere oovrecrol acpo aw, ai tfht ePr r2o0t e2c0t .eTd hAer pe ar os vdi edlei mr ai tl se od mlocaaplslyptroovcoidneddubcyt gCrFoIu2n0d1-t9r.uItfhdionugb. t persists, the provider may deploy `ield teams • Probabilistic risk of deforestation : The relevant third-party deforestation risk ap nr oa bl yasbi si l ipt ryoovfi df ue truarles od ec foonrde us tcat tsi oa nn faonrael yascihs ft aoramn?s”w. Tehr itshaenqaul yessitsi os nu :p“pWo rhtas t pi rs ot ahcet i v e rs iuspkpmo ri tt i. gUast ii no ng tbhyi si dme no dt i ef yl ,i nd ge ffoarrems tsatthi oant mc aany bree qpur ei rdei cetaerdl yboyuat pr epal yc ihn og ra asgpraotni aolmr ii sc k mp roodbealbtihl iat yt coof nf us itduer resdae vf oarrei es tt ya toi of ni nbdai cs ae tdoor sn aprroouxni md iet ay ctho pkleoyt . rWi s ke af ascsteosrss t. hTeh e s e iennccl ruodaec ht hme epnrte zs oe nn ec es , oafnfdo rhei ss tt owr ii tcha il nd ae fdoer `ei snteadt i ob nu fwf eirt, hwi nh iac hw mi d ae yr ri na dd ii cuas t, ewphoi ct ehnctai anl reveal patterns of nearby forest loss that signal ongoing or expanding threats. These
7
Made with FlippingBook flipbook maker