Using Remote Sensing to Quantify the Joint Effects of Climate and Land Use/Land Cover Changes on the Caatinga Biome of Northeast Brazilian

  Ci a ion: Ja dim, A.M.d.R.F.; A aújo Júnio , G.d.N.; Sil a, M.V.d.; San os, A.d.; Sil a, J.L.B.d.; Pando i, H.; Oli ei a-Júnio , J.F.d.; Teixei a, A.H.d.C.; Teodo o, P.E.; de Lima, J.L.M.P.; e al. Using Remo e Sensing o Quan i y he Join E ec s o Clima e and Land Use/Land Co e Changes on he Caa inga Biome o No heas B azilian. Remo e Sens. 2022,14, 1911. h ps://doi.o g/ 10.3390/ s14081911 Academic Edi o s: Baojie He, Ayyoob Sha i i, Chi Feng and Jun Yang Recei ed: 3 Ma ch 2022 Accep ed: 1 Ap il 2022 Published: 15 Ap il 2022 Publishe ’s No e: MDPI s ays neu al wi h ega d o ju isdic ional claims in published maps and ins i u ional a il- ia ions. Copy igh : © 2022 by he au ho s. Licensee MDPI, Basel, Swi ze land. This a icle is an open access a icle dis ibu ed unde he e ms and condi ions o he C ea i e Commons A ibu ion (CC BY) license (h ps:// c ea i ecommons.o g/licenses/by/ 4.0/). emo e sensing A icle Using Remo e Sensing o Quan i y he Join E ec s o Clima e and Land Use/Land Co e Changes on he Caa inga Biome o No heas B azilian Alexand e Maniçoba da Rosa Fe az Ja dim 1,2,*, Geo ge do Nascimen o A aújo Júnio 1,2, Ma cos Vinícius da Sil a 1, Ande son dos San os 1, Jhon Lennon Beze a da Sil a 1, Héli on Pando i 1, JoséF ancisco de Oli ei a-Júnio 3, An ônio He ibe o de Cas o Teixei a 4, Paulo Edua do Teodo o 5, João L. M. P. de Lima 6,7 , Ca los An onio da Sil a Junio 8, Luciana Sand a Bas os de Souza 2, Emanuel A aújo Sil a 9and Thie es Geo ge F ei e da Sil a 1,2 1Depa men o Ag icul u al Enginee ing, Fede al Ru al Uni e si y o Pe nambuco, Reci e 52171-900, B azil; [email p o ec ed] (G.d.N.A.J.); ma [email p o ec ed] (M.V.d.S.); [email p o ec ed] (A.d.S.); [email p o ec ed] (J.L.B.d.S.); [email p o ec ed] (H.P.); [email p o ec ed] (T.G.F.d.S.) 2Academic Uni o Se a Talhada, Fede al Ru al Uni e si y o Pe nambuco, Se a Talhada 56909-535, B azil; [email p o ec ed] 3Ins i u e o A mosphe ic Sciences, Fede al Uni e si y o Alagoas, Maceió57072-970, B azil; [email p o ec ed] 4Wa e Resou ces Depa men , Fede al Uni e si y o Se gipe, São C is ó ão 49100-000, B azil; [email p o ec ed] 5Depa men o Ag onomy, Fede al Uni e si y o Ma o G osso do Sul, Chapadão do Sul 79560-000, B azil; [email p o ec ed] 6MARE—Ma ine and En i onmen al Sciences Cen e, Uni e si y o Coimb a, 3000-456 Coimb a, Po ugal; [email p o ec ed] 7Depa men o Ci il Enginee ing, Facul y o Sciences and Technology, Uni e si y o Coimb a, 3030-788 Coimb a, Po ugal 8Depa men o Geog aphy, S a e Uni e si y o Ma o G osso (UNEMAT), Sinop 78555-000, B azil; [email p o ec ed] 9Depa men o Fo es Sciences, Fede al Ru al Uni e si y o Pe nambuco, Reci e 52171-900, B azil; [email p o ec ed] *Co espondence: alexand e.ja [email p o ec ed] Abs ac : Caa inga biome, loca ed in he B azilian semi-a id egion, is he mos populous semi-a id egion in he wo ld, causing in ensi ica ion in land deg ada ion and loss o biodi e si y o e ime. The main objec i e o his pape is o de e mine and analyze he changes in land co e and use, o e ime, on he biophysical pa ame e s in he Caa inga biome in he semi-a id egion o B azil using emo e sensing. Landsa -8 images we e used, along wi h he Su ace Ene gy Balance Algo i hm o Land (SEBAL) in he Google Ea h Engine pla o m, om 2013 o 2019, h ough spa io empo al modeling o ege a ion indices, i.e., lea a ea index (LAI) and ege a ion co e (V C ). Mo eo e , land su ace empe a u e (LST) and ac ual e apo anspi a ion (ET a ) in Pe olina, he semi-a id egion o B azil, was used. The p incipal componen analysis was used o selec desc ip i e a iables and mul iple eg ession analysis o p edic ET a . The esul s indica ed signi ican e ec s o land use and land co e changes on ene gy balances o e ime. In 2013, 70.2% o he s udy a ea was composed o Caa inga, while he lowes pe cen ages we e iden i ied in 2015 (67.8%) and 2017 (68.7%). Rain all eco ds in 2013 anged om 270 o 480 mm, wi h alues highe han 410 mm in 46.5% o he s udy a ea, concen a ed in he no he n pa o he municipali y. On he o he hand, in 2017 he lowes annual ain all alues ( om 200 o 340 mm) occu ed. Low ege a ion co e a e was obse ed by LAI and V C alues, wi h a ange o 0 o 25% ege a ion co e in 52.3% o he a ea, which exposes he e ec s o he d y season on ege a ion. The highes LST was mainly ound in u ban a eas and/o exposed soil. In 2013, 40.5% o he egion’s a ea had LST be ween 48.0 and 52.0 ◦ C, aising ET a a es (~4.7 mm day −1 ). Ou model has shown good ou comes in e ms o accu acy and conco dance (coe icien o de e mina ion = 0.98, oo mean squa e e o = 0.498, and Lin’s conco dance co ela ion Remo e Sens. 2022,14, 1911. h ps://doi.o g/10.3390/ s14081911 h ps://www.mdpi.com/jou nal/ emo esensing Remo e Sens. 2022,14, 1911 2 o 27 coe icien = 0.907). The signi ican inc ease in ag icul u al a eas has esul ed in he p og essi e educ ion o he Caa inga biome. The e o e, mi iga ion and sus ainable planning is i al o dec ease he impac s o an h opic ac ions. Keywo ds: opical d y o es ; su ace ene gy balance; B azilian semi-a id; SEBAL; ac ual e apo anspi a ion 1. In oduc ion The Caa inga biome occupies a la ge po ion o he B azilian semi-a id egion. I has a high ecological di e si y o plan species, and i is conside ed he la ges in he wo ld unde semi-a id condi ions [ 1 , 2 ]. I occupies an a ea o 900,000 km 2 , which co esponds o app oxima ely 70% o he no heas egion o B azil (NEB). Howe e , only 7.5% o his habi a is p o ec ed by law [ 3 – 5 ]. The municipali y o Pe olina, PE, B azil, is loca ed in he semi-a id egion o he Caa inga biome. I is wi hin he hyd og aphic basin o he São F ancisco i e , which a o s he de elopmen o i iga ed ag icul u e in he egion, mainly ui -g owing (e.g., g apes and mangoes) [ 6 – 8 ], and i leads o posi i e socioeconomic implica ions, bu also inc eases con lic s ega ding wa e use [2,9]. Caa inga is loca ed in he wo ld’s mos popula ed d y a ea, wi h mo e han 53 million inhabi an s and a popula ion densi y close o 34 inhabi an s pe km 2 . This biome has been a ec ed by en i onmen al deg ada ion and loss o biodi e si y o e he las decades, mainly by he in ensi ica ion o ag icul u e (e.g., ain ed and i iga ed c ops cul i a ion), u ban expansion, and he ad ance o pas u e a eas eplacing he na u al ege a ion [ 10 – 12 ]. Expanding ag icul u al ac i i ies has led o he de o es a ion o na i e a eas, soil dis u - bance, changes in he hyd ological cycle, and highe ca bon emissions [ 13 – 15 ]. Toge he wi h clima e changes, such as educed ain all and in ensi ied d ough e en s, his makes he Caa inga biome and he B azilian ecosys em he mos h ea ened and suscep ible o dese i ica ion. Fu he mo e, hese changes ha e been inc easingly comp omising na u al esou ces and en i onmen al sus ainabili y [ 16 – 19 ] due o he changes in su ace p ope ies and biophysical a iables, such as ege a ion co e (V C ), land su ace empe a u e (LST), and e apo anspi a ion (ET) [ 20 – 22 ]. ET is one o he main esponse pa ame e s o ege- a ed a eas as a unc ion o local wa e condi ions, and i is also an impo an componen o he hyd ological cycle [23,24]. The e o e, moni o ing physical–wa e indica o s o en i onmen al change condi ions, such as he loss o biodi e si y o biomes and land use and occupa ion, is i al in managing sca ce wa e esou ces. Fu he mo e, hese indica o s may be help ul in he planning o ag icul u al ac i i ies, as well as in he managemen o d y a eas and he sus ainable use and managemen o na u al esou ces [17,25–30]. In his scena io, emo e sensing has been used as a ool ha p esen s as and low ope a ional cos s, being e icien in calcula ing he biophysical pa ame e s used in he ene gy, wa e , and ege a ion balances (e.g., [ 31 – 33 ]). In ecen yea s, emo e sensing has also been conside ed a good al e na i e o eplacing expensi e and di icul - o-ob ain equipmen used in in si u s udies [ 34 ]. Fu he mo e, he modeling used in hese balances is pe o med wi h he help o algo i hms, which a e essen ial echniques in he ex ac ion o in o ma ion om sa elli e images on a egional and global scale [ 27 , 28 , 33 , 35 ]. I may ha e i s e iciency imp o ed by he use o open-sou ce cloud p og amming languages, e.g., Google Ea h Engine (GEE). In GEE, he use can e ec i ely implemen algo i hms and p ocess la ge da a olumes [36]. In his con ex , he e a e se e al su ace ene gy balance models, such as Mapping E apoT anspi a ion a high Resolu ion wi h In e nalized Calib a ion (METRIC), Su ace En- e gy Balance Sys em (SEBS), Simple Algo i hm o E apo anspi a ion Re ie ing (SAFER), A mosphe e–Land Exchange In e se (Alexi), and he Ene gy Balance Algo i hm o Land (SEBAL), based on emo e sensing da a. The app oach is subs an ia ed on biophysical pa ame e s, such as LST, albedo, No malized Di e ence Vege a ion Index (NDVI), and Remo e Sens. 2022,14, 1911 3 o 27 emissi i y, ha a e c ucial o es ima ing ET [ 16 , 37 – 42 ]. The SEBAL algo i hm has been widely and success ully applied o a ious wo ld ecosys ems, including semi-a id condi- ions in B azil [ 43 – 46 ]. Mo eo e , his me hod seeks o elimina e he p opaga ion o e o s in he pa i ioning o he ene gy balance and he need o a mosphe ic co ec ion in he es ima e o su ace empe a u e. These in e ac ions allow he gene a ion o he sensible hea lux co ec ed o a mosphe ic s abili y and ins abili y condi ions [ 47 – 51 ]. NDVI is one o he mos widely used indices in he li e a u e o ege a ion co e analysis, en i onmen al deg ada ion, and ege a ion p ima y p oduc ion esilience, and i also helps in moni o ing ag icul u al c ops, o es ecosys ems, and d ough assessmen s [ 52 – 55 ]. Wi h applica ions in se e al coun ies, Bas iaanssen e al. [ 49 ] ha e ob ained highly accu a e esul s using he SEBAL algo i hm on di e en ege a ed su aces wi h o es s, ag icul u al c ops (i iga ed and ain ed), and e en ex eme landscapes, such as dese a eas. P e ious s udies ha e shown ha he SEBAL can be used in Pe olina, PE, B azil, o he algo i hm has al eady been calib a ed and alida ed o he egion in a ious ecosys ems wi h good ag eemen be ween o bi al images and ield measu emen s [44,56–59]. Based on he abo e, he main objec i e o his pape is o de e mine and analyze he changes in land co e , land use, and occupa ion on biophysical pa ame e s in he Caa inga biome in he semi-a id egion o B azil. In he p esen s udy, he changes we e assessed om 2013 o 2019 using Landsa image y, and biophysical pa ame e s (ne adia ion, ene gy balance, LAI, V C , ET a , and LST) we e es ima ed by u ilizing emo e sensing. Addi ionally, he SEBAL model was applied o de e mine he su ace ene gy balance in di e en ege a ion en i onmen s. A e applying he model, he da ase p o ided by SEBAL allowed us o de e mine he u bulen luxes and ac ual e apo anspi a ion (ET a ) o he di e en land use and land co e (LULC) ypes. Thus, h ough he SEBAL p oduc s, he spa io empo al a ia ion pa e ns o ETain ag icul u al and o es y a eas we e e alua ed. 2. Ma e ials and Me hods 2.1. S udy A ea The s udy was ca ied ou in he municipali y o Pe olina, loca ed in he S a e o Pe nambuco, B azil. The egion comp ises he domain o he Caa inga biome and i belongs o he semi-a id egion o he sub-mean o he São F ancisco Valley. The own is conside ed he la ges ui -g owing cen e o he B azilian semi-a id egion due o he easy access o he São F ancisco i e , which supplies he i iga ed pe ime e s. The municipali y comp ises a e i o ial a ea o 4561.870 km 2 (Figu e 1), wi h an es ima ed popula ion o 349,145 inhabi an s [38]. A cha ac e is ic o he Caa inga biome is he p esence o di e en lo is ic mosaics, con- sis ing o an a ea o ee and sh ub ege a ion, which p esen s i s dis ibu ion condi ioned o clima ic and en i onmen al a ia ions, especially ain all in ensi y and equency [ 60 ], as well as geological con igu a ions and soil p ope ies [ 5 ]. The ege a ion o his biome p esen s deciduous species adap ed o wa e de ici condi ions and wi h exp essi e biomass p oduc ion in ainy seasons, esul ing om he local clima ic condi ions [ 40 , 61 ]. The canopy co e o he species o he Caa inga biome p esen s discon inuous cha ac e is ics, making possible he soil exposu e in d y pe iods, p esence o he baceous s a um, cac us species, and sh ubs [ 23 , 61 ]. I is wo h no ing ha his egion p esen s an exp essi e modi ica ion o he na i e landscape (Caa inga) in a eas o i iga ed ag icul u al cul i a ion. Remo e Sens. 2022,14, 1911 4 o 27 Figu e 1. Spa ial loca ion o he s udy a ea, municipali y o Pe olina, Pe nambuco, No heas B azil. Acco ding o he Köppen–Geige clima e classi ica ion, he egion’s clima e is o he BSh ype, cha ac e ized as semi-a id opical, wi h an a e age ai empe a u e o 26.4 ◦ C, a e age ela i e humidi y o 62%, and annual ain all o 520 mm [ 62 , 63 ]. Rain all pa e n is i egula h oughou he yea , esul ing om i s geog aphical loca ion and he In e op- ical Con e gence Zone (ITCZ) in luence, wi h ain all p edomina ing om Feb ua y o May [ 10 , 64 , 65 ]. The p edominan soils in he municipali y a e Typic Qua zipsammen , Ul isol Plin hic, A enosol, and Haplic Ac isol [66–68]. 2.2. Sa elli e Images and Wea he Da ase s Annual eco ds o a e age ai empe a u e ( ◦ C), global adia ion (MJ m −2 ), ela i e humidi y (%), a mosphe ic p essu e (kPa), wind speed (m s −1 ), and ain all (mm) we e ob ained om he da abase o he Na ional Ins i u e o Me eo ology [ 69 ] (Figu e 2). The a e age ai empe a u e (T a ) anged be ween 24.1 and 30.7 ◦ C, in which 2015 and 2019 we e he wa mes yea s s udied (T a 28 ◦ C). O e all, No embe , Decembe , Janua y, Feb ua y, and Ma ch p esen ed Ta alues abo e 28 ◦C (Figu e 2). The yea s 2013 (334.4 mm) and 2019 (221.6 mm) showed he highes ain all a es. Mos o he ain all eco ded o Pe olina was concen a ed om Decembe o Ma ch (Figu e 2). Such clima ic condi ions conce ning he municipali y we e also epo ed in he li e a u e (e.g., [2,70]). Rain all da a co esponding o he 30 days p io o he ou imaging da es s udied used we e ob ained om he Clima e Haza ds G oup In aRed P ecipi a ion wi h S a ion (CHIRPS). CHIRPS a e new p ecipi a ion p oduc s co e ing he coo dina es 50 ◦ S–50 ◦ N and 180 ◦ E–180 ◦ W, wi h 0.05 ◦ ( ± 5.3 km) spa ial esolu ion and daily o seasonal, empo al esolu ions, a ailable wo ldwide since 1981 [ 71 ]. CHIRPS da a we e ex ac ed om he Google Ea h Engine pla o m (h ps://ea hengine.google.com/, accessed on 20 Augus 2021) using Ja aSc ip p og amming language. Then, hey we e expo ed in sp eadshee o ma (*.xls), using he da ase since 1981 om he collec ion ee.ImageCollec ion (“UCSB- CHG/CHIRPS/DAILY”). Remo e Sens. 2022,14, 1911 5 o 27 Figu e 2. Mon hly me eo ological a ia ions ( ain all, a e age ai empe a u e, and global sola adia ion) o he municipali y o Pe olina, B azil, om 2013 o 2019. We used Ope a ional Land Image (OLI) Collec ion 1 Le el 1 bands 2 (0.450–0.51 µ m), 3 (0.53–0.59 µ m), and 4 (0.64–0.67 µ m) in he isible spec um, 5 (0.85–0.88 µ m) in he nea -in a ed, and 6 (1.57–1.65 µ m) and 7 (2.11–2.29 µ m) in he sho wa e in a ed, all wi h a spa ial esolu ion o 30 m, as well as band 10 om he The mal In a ed Senso (TIRS) wi h a 100 m spa ial esolu ion. Besides his, we used ou Landsa -8 OLI/TIRS images, pa h 217 and ow 66, co esponding o he yea s 2013, 2015, 2017, and 2019 ( o he da es and imes o he sa elli e o e pass, see Table 1). The choice c i e ia adop ed we e he absence o clouds (10%) and he images co esponding o he ansi ion pe iod be ween he d y and ainy seasons in he egion unde s udy, om he yea s o 2013 o 2019. This pe iod p esen ed se e e and ex eme d ough e en s in he no heas [ 23 , 72 ]. All he images we e ob ained om he Uni ed S a es Geological Su ey (USGS) pla o m (h ps://ea hexplo e .usgs.go /, accessed on 10 Augus 2021) and p ocessed h ough he Land Su ace Re lec ance Code (LaSRC). Table 1. Da e o he Landsa -8 sa elli e pass, ollowed by he Julian day (JD), Ea h–Sun dis ance (d , as onomical uni s—AU), local ime o he equa o pass (h, hou ; min, minu es), zeni h angle ( θ , ◦ ), sola ele a ion angle (E, ◦ ), and sun azimu h angle ( ϕ , ◦ ) o he municipali y o Pe olina, Pe nambuco, B azil. Acquisi ion Da e JD d Local Time θEϕ 5 Oc obe 2013 278 0.99 9 h 49 min a.m. 0.90 65.12 82.93 12 No embe 2015 316 0.99 9 h 48 min a.m. 0.90 64.85 113.46 16 Oc obe 2017 289 0.99 9 h 48 min a.m. 0.91 65.81 92.82 7 No embe 2019 311 0.99 9 h 48 min a.m. 0.90 65.41 110.36 No e: zeni h angle (θ) = sin(E). Sou ce: USGS/NASA [73]. 2.3. Vege a ion Indices The No malized Di e ence Vege a ion Index (NDVI) was calcula ed o ep esen he amoun and quali y o ege a ion p esen on he su ace, cha ac e ized as an indica o o we condi ions, calcula ed using Equa ion (1). NDVI =ρNIR −ρRed ρNIR+ρRed (1) whe e ρNIR and ρRed a e he e lec ances measu ed in he nea -in a ed and ed bands (i.e., Landsa -8 mul ispec al bands 5 and 4 o he OLI senso ), espec i ely, hey ange om − 1 o +1. Values close o 1 on a posi i e scale co espond o high pho osyn he ic ac i i y, and when nega i e, gene ally co espond o wa e bodies. Remo e Sens. 2022,14, 1911 6 o 27 Based on he NDVI, we calcula ed he ege a ion co e (V C ) o he s udy a ea (Equa ion (2)) , acco ding o Gao e al. [54]. VC=NDVI −NDVIS NDVIV−NDVIS ·100 (2) whe e V C is he ege a ion co e , NDVI S is he minimum NDVI alue om ba e soil pixels ob ained in he s udy a ea, and NDVI V is he maximum NDVI alue ound in ege a ed a eas, i.e., om ully ege a ed pixels. The NDVI S and NDVI V used o calcula e V C we e ob ained om he domain o each NDVI image pe cen ile map ob ained om he NDVI his og ams. Soil-Adjus ed Vege a ion Index (SAVI) was calcula ed o obse e he ege a ion co e o he a ea (Equa ion (3)). SAVI =(1+L)·(ρNIR −ρRed) (L+ρNIR+ρRed)(3) whe e Lis he adjus men ac o o he soil, which a ies be ween 0 and 1. The alue 0 does no each change, and esembles he NDVI. In a eas wi h low-densi y ege a ion, he alue 1 is assigned; o in e media e-densi y ege a ion a eas, he alue o 0.5; and o a eas wi h high-densi y ege a ion, he alue 0.25 is assigned [ 74 ]. The adjus men ac o o 0.5 was adop ed due o he s udy egion indica ing an in e media e ege a ion co e age in mos o he yea , wi h p edominan ege a ion o he Caa inga biome, in he B azilian semi-a id egion [75–77]. To e alua e changes in ege a ion biomass, he lea a ea index (LAI, m 2 m −2 ) was de e mined (Equa ion (4)), a undamen al biophysical a iable o moni o ing s udies o ag icul u al land and ege a ion mois u e condi ions [50]. LAI = −ln0.69−SAVI 0.59  0.91 (4) 2.4. Me hodology o Es ima ing E apo anspi a ion Using Sa elli e Images E apo anspi a ion was es ima ed using he Ene gy Balance Algo i hm o Land (SEBAL). Fo his, ou ine me eo ological da a and spec al bands om he Landsa -8 sa elli e we e used. The SEBAL algo i hm was implemen ed using Ja aSc ip code h ough he Google Ea h Engine (GEE) pla o m. The da a we e expo ed in sp eadshee o ma (*.xls). SEBAL uses ma hema ical modeling and ope a ions o calcula e he su ace ene gy balance componen s and de e mine e apo anspi a ion. Thus, ene gy balance componen s a e compu ed pixel-by-pixel, as desc ibed in Figu e 3. He e, we show ha he algo i hm has a good pe o mance and high accu acy, as well as being calib a ed and alida ed wi h simul aneous ield and Landsa sa elli e measu emen s [47–49,56,59,61,78,79]. Remo e Sens. 2022,14, 1911 7 o 27 Figu e 3. Flowcha o he SEBAL model o es ima ing e apo anspi a ion. No e: LST and LSE a e he land su ace empe a u e and land su ace emissi i y, espec i ely, NDVI is he No malized Di e ence Vege a ion Index, εa is he a mosphe ic emissi i y, u* is he ic ion eloci y, T a is he ai empe a u e, R n is he ne adia ion, Gis he soil hea lux, z om is he momen um oughness leng h, z 1 and z 2 a e he wo heigh s be ween he su ace o he ancho pixels, ah is he nea -su ace ae odynamic esis ance o hea anspo , DEM is he digi al ele a ion model, dT is he nea -su ace ai empe a u e g adien , aand ba e he calib a ion coe icien s, LST o T s is he land su ace empe a u e, ρai is he ai densi y, C p is he speci ic hea o ai , His he sensible hea lux, ψm and ψh a e he s abili y co ec ion ac o s o momen um and sensible hea , espec i ely, kis he on Ka man cons an , u 200 is he wind speed a he heigh o 200 m, and Λ is he e apo a i e ac ion. These p e- and p ocessing s eps we e pe o med inside Google Ea h Engine (GEE) cloud pla o m. The ac onyms and symbols used in his s udy a e summa ized in he Abb e ia ions sec ion. Su ace Albedo Adjus men The su ace albedo ( αsup ) co esponds o a measu e o he e lec i i y o he Ea h’s su ace, o each pixel, wi h a mosphe ic co ec ion ob ained acco ding o Equa ion (5) [ 50 , 80 ]. αsup =α oa −αpa h τsw2(5) whe e α oa is he albedo a he op o he a mosphe e, ha is, be o e a mosphe ic co ec ion, αpa h is he a mosphe ic e lec ance (se o 0.03, as used by Sil a e al. [ 80 ]), and τsw is he a mosphe ic ansmissi i y o clea sky condi ions, acco ding o Equa ion (6) [50,80]: τsw=0.35 +0.627 ·exp"−0.00146 ·Pa K ·cos(θ)−0.075W cos(θ)0.4#(6) whe e P a is he a mosphe ic p essu e (kPa), wi h da ase a ailable eely in h ps://po al. inme .go .b / (accessed on 10 Augus 2021), K is he u bidi y coe icien o he a mosphe e (K = 1.0, o a clea sky day), acco ding o Allen e al. [ 47 ] and Sil a e al. [ 80 ], θ is he sola zeni h angle, and Wis he p ecipi able wa e (mm), es ima ed om Equa ion (7) [81]. W=0.14 ·ea·Pa+2.1 (7) whe e e a is he ac ual a mosphe ic wa e apo p essu e (kPa), es ima ed om Equa ion (8). ea=HR ·es 100 (8) Remo e Sens. 2022,14, 1911 8 o 27 whe e HR is he ins an aneous ela i e humidi y (%), and e s is he wa e apo sa u a ion p essu e (kPa), es ima ed om Equa ion (9). es=0.6108 ·exp17.27 ·T0 237.3 +T0(9) whe e T0is he ins an aneous ai empe a u e (◦C) a he momen o he sa elli e pass. Fo ob aining α oa , a linea combina ion o he spec al e lec ance o he six e lec i e OLI bands was pe o med acco ding o Equa ion (10) [80]: α oa =0.300 2+0.277 3+0.233 4+0.143 5+0.036 6+0.001 7(10) whe e 2 , 3, 4 , 5, 6 , and 7 a e he su ace spec al e lec ances o bands 2, 3, 4, 5, 6, and 7 o he Landsa -8 OLI, espec i ely. We use Equa ion (11) o ob ain each o he spec al e lec ances. b=Addb+Mul b·DN cos(θ)·d (11) whe e he e ms Add b and Mul b belong o he adiome ic escaling g oup, speci ically e lec ance_add_band (equal o − 0.1) and e lec ance_mul _band (equal o 0.00002), espec- i ely, p esen ed in he me ada a o each OLI—Landsa -8 image, DN is he digi al numbe alue co esponding o he pixel, θ is he sola zeni h angle a he da a acquisi ion ime, and d is he Ea h–Sun dis ance in as onomical uni s. 2.5. De e mina ion o Su ace-ene gy Pa i ioning Based on he su ace ene gy balance componen s, he e apo a i e ac ion was de- e mined. Ini ially, he su ace adia ion balance o ne adia ion—R n was calcula ed, which is dis ibu ed by he ene gy pa i ioning in on o he sensible hea luxes—H, la en —LE, and soil hea lux—G[ 47 – 50 , 61 , 78 , 79 ]. By pe o ming his p ocess, a linea ela ionship be ween he su ace and ai empe a u e g adien was conside ed o exis . F om his ela ionship and he in e nal calib a ion p ocess o ex eme condi ions such as empe a u e and humidi y, i was es ablished he need o ob aining he knowledge o he so-called “ancho pixels”, i.e., ho and cold pixels, which a e indica i e o ze o and maximum e apo anspi a ion, espec i ely [47,48,61] (Figu e 3). The land su ace empe a u e (LST) in K (Kel in) was ob ained using he spec al adiance in band 10 o he TIRS senso and he emissi i y in he nea es band— εnb by he modi ied Planck’s Law [82], as desc ibed in Equa ion (12). LST =K2 ln” nb·K1 L10 +1(12) whe e K 1 and K 2 a e adia ion cons an s speci ic o he Landsa -8 TIRS band 10, equaling 774.89 W m−2s −1µm−1and 1321.08 K, espec i ely, p o ided by NASA/USGS; and L10 is he adiance a he wa eleng h ecei ed by he senso s (band 10, he he mal band). The εnb was calcula ed based on he LAI o each pixel acco ding o Equa ion (13) [ 41 ]. ” nb=0.97 +0.0033 ·LAI (13) Ini ially, he empe a u e a ia ion and ae odynamic esis ance o hea anspo in all pixels o he s udy a ea (Pe olina, Pe nambuco) we e de e mined. The a mosphe e was ini ially assumed o be in a neu al s abili y condi ion. Fo his s udy, he ho pixel was conside ed in he exposed soil plo s (i.e., no ege a ion co e and/o li le ege a ion and low mois u e con en ), assuming LE equal o ze o. The cold pixel was conside ed in g ape o cha d plo s i iga ed by mic o-sp inkle s, when Hcan be conside ed ze o [ 47 , 48 , 61 , 78 , 79 , 83 ] (see Figu e 3). Since u bulen e ec s a ec a mosphe ic condi ions and ai esis ance, he Remo e Sens. 2022,14, 1911 9 o 27 Monin–Obukho simila i y heo y was applied and conside ed in he compu a ion o Hin all pixels o he s udy a ea. I is wo h no ing ha he Monin–Obukho leng h was used o co ec ions o he ini ial s able condi ion o he a mosphe e [47–49,78,79]. Calcula ion o Ene gy Fluxes (Hand LE) and E apo a i e F ac ion The sensible hea lux (H) in SEBAL is calcula ed using an i e a i e p ocedu e om he ae odynamic unc ion (Equa ion (14)) [47–49,78,79]. H=ρai ·Cp·(a+b·LST) ah (14) whe e ρai is he mois ai densi y (kg m −3 ), C p is he ai speci ic hea a cons an p es- su e (1004 J kg −1 K −1 ), aand ba e calib a ion cons an s o he empe a u e di e ence be ween wo heigh s (i.e., be ween he oughness leng h o hea ans e and he e e ence heigh , usually 0.1 and 2.0 m abo e he displacemen plane), and ah is he nea -su ace ae odynamic esis ance o hea anspo (s m −1 ). Fundamen ally, he coe icien s aand ba e de e mined h ough an in e nal calib a ion o each sa elli e image by in e ac i e p ocesses. We conside ex eme pixels o we /cold and d y/ho spo s. They we e selec ed o de elop a linea ela ionship be ween he ae odynamic empe a u e o he su ace and he ai empe a u e di e ence, and he LST. By knowing he componen s o he su ace ene gy balance, such as he ne adia ion (R n ,Wm −2 ), sensible hea lux (H,Wm −2 ), and soil hea lux (G,Wm −2 ), he la en hea lux (LE,Wm −2 ) was de e mined, bo h co esponding o he ime o he sa elli e pass o e he s udy a ea, acco ding o Equa ion (15). LE =Rn−H−G(15) Subsequen ly, we de e mined he e apo a i e ac ion ( Λ ) acco ding o Equa ion (16). Λ=LE Rn−G(16) 2.6. Es ima e o ETaUsing SEBAL Me hod Finally, as a SEBAL p oduc , we de e mine he ac ual e apo anspi a ion (ET a , mm day −1 ) [ 84 ] based on Equa ion (17) below, o each sa elli e image used. In his s udy, he implemen ed SEBAL model had al eady been ex ensi ely alida ed and calib a ed unde o es s and ag icul u al land condi ions [56,61,85]. ETa=Λ·Rn24 ·86, 400 ˘(17) whe e R n24 is he daily ne adia ion (W m −2 ), 86,400 is a cons an o daily imescale con e sion (i.e., con e s om seconds o days), and λ is he la en hea o apo iza ion o wa e (J kg −1 ). Then, he la en hea o apo iza ion allows he ET a exp ession in mm day −1 . Hence, accu a e es ima ion o R n24 (Equa ion (18)) was de e mined acco ding o Bas iaanssen e al. [49], and Lee and Kim [86]: Rn24=Λ·(1−αsup)·Rn−a·τsw(18) whe e ais a eg ession coe icien o he ela ionship be ween ne longwa e adia ion and a mosphe ic ansmissi i y on a daily scale, o which we assigned he alue 143, as p oposed by Teixei a e al. [ 61 ]. The ac onyms and symbols used in his s udy a e summa ized in he Abb e ia ions sec ion. Remo e Sens. 2022,14, 1911 16 o 27 dis ibu ion o ain all accumula ion in he p e ious days. On he o he hand, du ing he da es s udied he e, he highes mean LAI alues s ood ou only in a eas o a bo eal Caa inga (0.58 ± 0.45 m 2 m −2 ) and ag icul u e (0.75 ± 0.49 m 2 m −2 ), wi h he maximum alues being associa ed wi h i iga ed ag icul u al a eas, mo e speci ically o cha ds, wi h inc eased biomass p oduc ion. Howe e , in a eas o pas u e, u ban in as uc u e, and a eas o a bo eal and he baceous Caa inga, he LAI alues we e close o o equal o ze o, making i s a ia ion mo e homogeneous, ha is, close o he daily mean, wi h s anda d de ia ion (SD) anging be ween 0.05 and 0.10 m2m−2wi hin he land co e classes (Figu e 7). 3.4. Land Su ace Tempe a u e (LST) in he S udied Classes In he p esen s udy, he minimum LST alues we e seen in a eas o wa e bodies, while he maximum LST alues we e seen in a eas o exposed soils, loca ed a poin s o pas u e and deg aded Caa inga, u ban in as uc u e (asphal , conc e e, and g a el su aces), and ag icul u al a eas unde going soil p epa a ion o cul i a ion (Figu e 8). This esul is expec ed in ba e soil loca ions unde in ense an h opic ac i i y due o he ans o ma ion o land use/land co e classes in o non-e apo a ing su aces. This makes he place’s empe a u e highe and educes wa e a ailabili y in he soil, which causes se ious p oblems in ag icul u al c ops. The compu ed LST map is shown in Figu e 8. Figu e 8. Spa io empo al dis ibu ion o he land su ace empe a u e—LST ( ◦ C) in he municipali y o Pe olina, Pe nambuco, B azil, on he imaging da es 5 Oc obe 2013 ( a ), 12 No embe 2015 ( b ), 16 Oc obe 2017 (c) and 7 No embe 2019 (d). Due o he eplacemen o p ima y ege a ion wi h pas u es, ag icul u al c ops, and u ban occupa ion, changes in land use can subs an ially a ec he hea and mass exchange in he soil–plan –a mosphe e sys em, p opi ia ing he e en ion o a highe amoun o hea by he Ea h’s su ace [ 2 , 52 ]. Land abandonmen and excessi e mechanical dis u bance o he soil may also al e he hea exchange wi h he en i onmen and cause lowe he mal and adian ene gy lag; hus, he land con e sion had inc eased LST in he a ea o he non-e apo a ing su aces. I can be obse ed ha he a e age LST alues o he da es s udied we e highe in he a eas domina ed by pas u e (47.69 ± 1.47 ◦ C), he baceous Caa inga (47.28 ± 1.27 ◦ C), and sh ub Caa inga (46.07 ± 1.44 ◦ C), e en highe han hose obse ed in a eas wi h u ban Remo e Sens. 2022,14, 1911 17 o 27 in as uc u e (45.80 ± 1.52 ◦ C) (Figu e 8). Acco ding o Zhao e al. [ 110 , 111 ], si es wi h di e en land co e ypes may ha e an LST inc ease g adien along he u ban o u al p o ile. The high LST in he pas u e, he baceous, and sh ub Caa inga a eas is ela ed o he lowe pe cen age o g ound co e ing by ege a ion (see Figu e 6), which esul s in d ie exposed soil, wi h highe albedos and lowe e apo a i e cooling lux a es, a ac o ha inc eases LST. Ano he ela ed ac o con ibu ing o he high LST o pas u es is ha g asses ha e shallowe oo s. The e o e, hey can only access he wa e a ailable in he supe icial soil laye s, which deple es as e han in deepe laye s [ 2 ]. Vege a ion canopy can e ain ainwa e and dec ease g oundwa e echa ge by al e ing e apo a i e lux and aising he land su ace empe a u e. On he o he hand, i is obse ed, in gene al, ha in a eas domina ed by a bo- eal Caa inga and ag icul u e, he a e age LST alues a e lowe (38.99 ± 2.48 ◦ C and 43.11 ± 2.30 ◦ C, espec i ely) (Figu e 8). The highes ege a ion co e and he highes soil humidi y in he a eas domina ed by hese classes a o LST educ ion. Howe e , hey p esen he g ea es spa ial a ia ions o LST among all land use and land co e classes, acco ding o he s anda d de ia ion alues ( ± SD). The main ad an age o using LST da a om sa elli e images is he o al su ace co e age. In his way, each ime se ies o pixels o he LST map can be conside ed a “ i ual wea he s a ion” [112]. 3.5. Va ia ions o he Ac ual E apo anspi a ion (ETa) o Land Use Classes Rain all egime di ec ly in luenced ET a , so ha on 16 Oc obe 2017 (Figu e 9c), he da e wi h he lowes ain all accumula ion in he p e ious days, he lowes ET a a es we e ound, wi h an a e age alue o 2.02 mm day −1 . On he o he hand, on 7 No embe 2019 (Figu e 9d), he pe iod wi h he highes ain all accumula ion, a e age ET a a es we e 2.62 mm day −1 (Figu e 9). Acco ding o Teixei a e al. [ 20 ], high e apo anspi a ion alues in Caa inga a eas occu igh a e ains. The e o e, he p e ious ain all aises soil wa e a ailabili y and keeps na i e species wi h u gid s uc u es and g eene canopy. Figu e 9. Spa io empo al dis ibu ion o ac ual e apo anspi a ion (ET a , mm day −1 ), calcula ed wi h Su ace Ene gy Balance Algo i hm o Land (SEBAL), in he municipali y o Pe olina, Pe nambuco, B azil, on he imaging da es 5 Oc obe 2013 ( a ), 12 No embe 2015 ( b ), 16 Oc obe 2017 ( c ), and 7 No embe 2019 (d). Remo e Sens. 2022,14, 1911 18 o 27 The ET a es ima ed by he SEBAL model showed a ia ion bo h wi hin and be ween land use and land co e classes. The lowes ET a obse a ions in all he e alua ed da es we e obse ed in he a eas occupied by pas u e and mosaic o ag icul u e and pas u e classes, wi h a e age alues o 0.70 ± 0.73 mm day −1 and 1.01 ± 0.96 mm day −1 , espec i ely (Figu e 9). These a eas ha e d yland cul i a ion p ac ices, which causes he lowe wa e a ailabili y o a ec he e apo anspi a ion a es; mo eo e , he he e ogenei y o he a eas causes sudden a ia ions in ETa. On he o he hand, due o he e ec o he inc ease in ai empe a u e and a mosphe ic demand e i ied h oughou he d y season in Pe olina, combined wi h he p esence o p ese ed ipa ian o es s along s e ches o wa e bodies and con inuous i iga ion in c ops, highe mean ET a alues we e obse ed in he a bo eal Caa inga (4.73 ± 0.49 mm day −1 ) and ag icul u e (3.07 ± 1.23 mm day −1 ) classes. Fo p esen ing a eas wi h i iga ed and d y cul i a ion, he a e age alues o his class become mo e a iable. When he e is a g ea e con ibu ion o mois u e added o mo e dense ege a ion, he e is a a o ing o he local mic oclima e in he egion [113,114], a phenomenon epo ed in a eas o a bo eal ege a ion. The sh ub and he baceous Caa inga classes showed g ea e he e ogenei y indica ed by he la ges s anda d de ia ions (2.42 ± 0.76 mm day −1 ) and (1.46 ± 0.71 mm day −1 ), espec i ely, ela i e o he a bo eal Caa inga class (Figu e 9). Folhes e al. [ 115 ] epo ed ha he e apo anspi a ion alues (2.0 mm day −1 ) du ing he d y season o species o he baceous–sh ubby Caa inga. In d y pe iods, he Caa inga ege a ion uses he a ailable ene gy as sensible hea lux (H), limi ing anspi a ion and pho osyn hesis, hus educing e apo anspi a ion alues [ 20 ]. Howe e , a bo eal ege a ion is able o compensa e he high apo p essu e de ici in he ai , e en in d y pe iods, when compa ed o sh ub and he baceous ege a ion, due o he deep oo sys em keeping up wi h he wa e s o ed in he soil [70,116,117]. The a bo eal and sh ub species play a undamen al eco-hyd ological ole, main aining soil humidi y and s uc u ing i s po osi y, gua an eeing he main enance o in il a ion capaci y and a o ing he su i al o species [116]. 3.6. S a is ical Rela ions be ween he Va iables S udied and Land Use In his s udy, we pe o med a PCA o he en i onmen al a iables in ela ion o land use and land co e classes. The e o e, he i s wo componen s wi h eigen alues g ea e han 1.0 we e ex ac ed sepa a ely o he yea s 2013, 2015, 2017, and 2019 (Figu e 10). In 2013, he wo p incipal componen s explained 94.77% o he o al a ia ion, wi h 70.39% in he p incipal componen 1 (PC1) and 24.38% in he p incipal componen 2 (PC2). On he o he hand, in 2015, 2017, and 2019, when added oge he , PC1 and PC2 ep esen ed 94.01, 92.34, and 94.56% o he o al a ia ion, espec i ely (Figu e 10). In addi ion, i can be seen ha he LULC class, wi h he leas in luence on componen s 1 and 2 in all yea s, is sh ub Caa inga, wi h a e age eigen alues (0.31 and 0.37, espec i ely). In addi ion, he classes wi h he g ea es in luence on PC1 wi h posi i e and nega i e eigen alues a e wa e bodies (2.69), ag icul u e ( − 2.35), a bo eal Caa inga ( − 1.92), pas u e (0.34), mosaic (0.16), and u ban a ea (0.35). In PC2, he classes o LULC we e wa e bodies ( − 4.36), mosaic (2.10), ag icul u e (−1.0), and a bo eal Caa inga (−3.40) (Figu e 10). Remo e Sens. 2022,14, 1911 19 o 27 Figu e 10. Sco es ob ained by p incipal componen analysis (PCA) o en i onmen al a iables and land use and land co e . PC1 and PC2 a e he i s and second dimensions o PCA da a, espec i ely. The ou inse ed panels below he PCA sco es plo s e e o he loadings plo s o he i s wo p incipal componen s om 2013 o 2019. Th ough PCA, we obse ed ha he o de ing o a iables in each p incipal componen (PC) o he axes was in luenced by he deg ee o ege a ion co e and su ace wa e s a us o he LULC classes (Figu e 10). Thus, PC1 con ibu ed mo e o he a iabili y o he esponse o a iables ela ed o ene gy balance. The e o e, in PC2, he land use and land co e classes (i.e., pas u e, mosaic, u ban a ea, and he baceous Caa inga) in luenced he a iables R n ,LE, ET a , and emissi i y wi h highe mean loadings ( − 0.95, − 0.94, − 0.99, and − 0.80, espec i ely). On he o he hand, hey showed a high co ela ion wi h he a iables H, LST, and albedo. Fo hese LULC classes, his may be ela ed o he p esence o ba e soil and hin ege a ion co e s ha a ec he egional mic oclima e and soil–plan –a mosphe e sys em luxes, esul ing in highe albedos, lowe a es o e apo a i e cooling luxes, and highes LST [2,105]. On he o he hand, PC2 con ibu ed mo e o he a iabili y o he esponses o he a iables ega ding he canopy in e ac ions (LAI and V C ) due o he s ong nega i e co ela ion wi h cul i a ed land (i.e., ag icul u e class) (Figu e 10). In all yea s, he e was a p edominance o LE and LAI in a eas wi h a bo eal Caa inga and ag icul u e, wi h ag icul u e p esen ing he highes V C . Caa inga p esen s a s ong ela ionship wi h LAI in ainy pe iods due o he g ea e a ailabili y o wa e in he soil [ 118 ]. In he p esen s udy, he samples we e aken in he pe iod wi h low ain all, so he LAI was no exp essi e compa ed o ag icul u al a eas, which use i iga ion and he e o e inc ease he LAI. Thus, Remo e Sens. 2022,14, 1911 20 o 27 he emissi i y in he Caa inga ege a ion is lowe , o he emissi i y o he soil is gene ally lowe han ha o he lea es [2]. The esul s o he analysis o a iance (ANOVA) and mul iple eg ession analysis o he es ablished model a e p esen ed in Table 3. Fo he combina ion o wo- a iable models, he wo bes p e-es ablished pa ame e s based on he PCA esul s we e he a iables LST and H, used o build he eg ession model. In pa icula , hese wo a iables a e o g ea ele ance in ans e ing ene gy o he a mosphe e. The ANOVA esul s also showed ha he model alues a e signi ican . Due o he obse a ions, a join analysis o he coe icien o de e mina ion ob ained (R 2 = 0.98) can be pe o med, emphasizing he P- alue ob ained om ou eg ession model, which was less han 0.001, hus indica ing g ea e model accu acy and eliabili y (Table 3). No ably, i can be seen ha he model’s F- alue was 16,692.84, being g ea e han he c i ical alue o F 0.05 = 3.018, which con i ms he signi icance o he p oposed model. In addi ion, he a iables used p o ided a high R 2 and LCCC, being essen ial o he model’s accu acy. Fu he mo e, he esul s showed ha he mul iple linea eg ession model o de e mine ET a achie ed a coe icien o de e mina ion o 0.98, RMSE o 0.498, MAE o 0.413, and dequal o 0.9620. I also esul ed in PBIAS, NSE, and LCCC alues a e aged be ween − 13.32%, 0.826, and 0.907, espec i ely (Table 3). F om he s a is ical analysis, his model is desc ibed as ET a = 6.89 − 0.0527LST − 0.0120H. Based on he RMSE, LCCC, and do his applica ion, he use o he ET a model, besides p esen ing a s ong co ela ion be ween he a iables Hand LST, as seen in Figu e 10, exp esses a biophysical model wi h enough e iciency and high ag eemen o de e mine ETa. Table 3. Analysis o a iance (ANOVA) and eg ession coe icien s esul s o he sugges ed model. Sou ce o Va ia ion d SS MS F-Value p-Value Reg ession 2 464.41 232.21 16,692.84 0.0001 LST 1 335.29 335.29 24,103.8 0.0001 H1 129.12 129.12 9281.9 0.0001 E o 397 5.52 0.01 To al 399 469.93 Reg ession s a is ics P edic o s in model Reg ession coe icien s β0β1β2R2 LST, H6.89 −0.0527 −0.0120 0.98 Model S a is ical me ics RMSE MAE PBIAS (%) NSE LCCC d 0.498 0.413 −13.32 0.826 0.907 0.9620 d : deg ees o eedom, SS: sum o squa es, MS: mean squa e, LST: land su ace empe a u e, H: sensible hea lux, β0 : in e cep , β1 and β2 : es ima ed coe icien o he ac o x, R 2 : coe icien o de e mina ion, RMSE: oo mean squa e e o , MAE: mean absolu e e o , PBIAS: pe cen bias, NSE: Nash–Su cli e e iciency coe icien , LCCC: Lin’s conco dance co ela ion coe icien , and d: Willmo ’s index o ag eemen . Based on F- es , a a p obabili y o 0.05 (p< 0.05), signi icance o equa ion pa ame e s o each esponse a iable was de e mined. The high d(0.9620) alues o ET a indica ed ha he e was a good ag eemen be ween simula ed and measu ed ET a . In gene al, he ET a model showed excellen ag eemen (LCCC 0.9), high pe o mance, and low RMSE (0.498) (Table 3). The PBIAS and MAE alues o ET a we e be ween − 13.32% and 0.413, con i ming he close ag eemen . Consequen ly, his esul desc ibes he abili y o simpli y and accu a ely p edic ET a in de ici en i onmen s. Applica ions o eg ession model analysis wi h en i onmen al a iables a e common in he li e a u e in d y o es s [ 118 – 120 ]. Howe e , hese implemen a ions o a iables in p e ious s udies may be challenging o acqui e o speci ic loca ions, equi ing mo e simpli ied models. Ou esul s also indica e he ela ionship be ween Hand he LST o ecosys ems o de e mine ETa, an impo an a iable in he ene gy balances o en i onmen s wo ldwide. Remo e Sens. 2022,14, 1911 21 o 27 4. Conclusions In his s udy, we poin ou a signi ican endency o inc ease he ag icul u al a eas, which esul s in he p og essi e dec ease o he B azilian Caa inga biome. The ege a ion co e is di ec ly in luenced by he soil–wa e egime; yea s o highe ain all esul in a lowe pe cen age o supp ession o he na i e o es in he municipali y o Pe olina, Pe nambuco (B azil). The a eas wi h pas u e class p esen ed ho spo s due o deg ada i e p ocesses and highe su ace empe a u es, in luenced by he sensible hea lux. A g adual inc ease in LST is obse ed in he municipali y and i may cause u u e isks o o es a eas. The SEBAL algo i hm used in a semi-a id en i onmen is a help ul ool o de e mine he ene gy and mass luxes in di e en ecosys ems. No ably, he Caa inga biome has pa icula i ies in biophysical pa ame e s, acco ding o he land co e and soil exposu e on in a and in e -annual scales. The he e ogenei y o he su ace o he municipali y o Pe olina, as a unc ion o land use and land co e pa e ns, al e s he ene gy exchange wi h he a mosphe e. Ou esul s also sugges a simpli ied and alida ed model o ET a de e mina ion in a semi-a id en i onmen . The eg ession model could accu a ely p edic he spa ial dis ibu ion o ETa, wi h high R2and LCCC and low RMSE alue. Thus, i is possible o sugges ha he implemen a ion o ag icul u al ac i i ies in he Pe olina should be ca ied ou in a planned and sus ainable way in o de o mi iga e he impac s ha an h opic ac ion causes on he Caa inga, especially wi h he inc eased ulne abili y o his biome o he dese i ica ion p ocess. Howe e , u he esea ch is needed o in es iga e he spa ial a ia ions o he ypes o c ops co e ing he soil in he municipali y, as well as he dynamics o i es and hei impac s on he di e si y o he Caa inga biome. Field su eys and he use o unmanned ae ial sys ems (UAS) could p o ide mo e de ailed in o ma ion a an in e media e and ine scale. Supplemen a y Ma e ials: The ollowing a e a ailable online a h ps://www.mdpi.com/a icle/ 10.3390/ s14081911/s1, Figu e S1: Land use/land co e changes in Pe olina be ween 2013 and 2019. Posi i e alues indica e an expansion o he espec i e land co e , nega i e alues a con ac ion. The e ical axe is in million hec a es (Mha). Please see Supplemen a y Table S1 o access alues o land use/land co e changes; Table S1: A eas o expansion and con ac ion o land use/land co e changes in Pe olina be ween 2013 and 2019. Posi i e alues indica e an expansion o he espec i e land co e , nega i e alues a con ac ion. The alues a e in million hec a es (Mha). Au ho Con ibu ions: Concep ualiza ion, A.M.d.R.F.J. and G.d.N.A.J.; me hodology, M.V.d.S., A.d.S., and A.M.d.R.F.J.; so wa e, M.V.d.S., A.d.S., and A.M.d.R.F.J.; alida ion, J.F.d.O.-J. and A.H.d.C.T.; in es iga ion, J.L.B.d.S., H.P., P.E.T., L.S.B.d.S., and C.A.d.S.J.; da a cu a ion, M.V.d.S., A.d.S., A.M.d.R.F.J., and G.d.N.A.J.; w i ing—o iginal d a p epa a ion, A.M.d.R.F.J.; w i ing— e iew and edi ing, T.G.F.d.S., J.L.M.P.d.L., and E.A.S.; isualiza ion, J.L.M.P.d.L. and A.H.d.C.T.; supe ision, A.H.d.C.T., T.G.F.d.S., and J.F.d.O.-J.; p ojec adminis a ion, T.G.F.d.S. and J.L.M.P.d.L.; unding acquisi ion, J.L.M.P.d.L. All au ho s ha e ead and ag eed o he published e sion o he manusc ip . Funding: This esea ch was unded by he Po uguese Founda ion o Science and Technology (FCT), h ough p ojec s ASHMOB (CENTRO-01-0145-FEDER-029351), GOLis (PDR2020-101-030913, Pa - ne ship n . 344/Ini ia i e n . 21), MUSSELFLOW (PTDC/BIA-EVL/29199/2017), MEDWATERICE (PRIMA/0006/2018), and h ough he s a egic p ojec UIDB/04292/2020 g an ed o MARE—Ma ine and En i onmen al Sciences Cen e, Uni e si y o Coimb a, Coimb a, Po ugal. Ins i u ional Re iew Boa d S a emen : No applicable. In o med Consen S a emen : No applicable. Da a A ailabili y S a emen : Landsa -8 image cou esy o he USGS/NASA (h ps://www.usgs. go /co e-science-sys ems/nli/landsa , accessed on 10 Augus 2021); MapBiomas da a p esen ed in his s udy a e a ailable a websi es o B azilian Annual Land Use and Land Co e Mapping P ojec (h ps://mapbiomas.o g/en/p ojec , accessed on 20 Augus 2021); and he me eo ological da a p esen ed in his s udy a e a ailable a he websi e o he Na ional Ins i u e o Me eo ology (h ps://po al.inme .go .b /, accessed on 10 Augus 2021). Remo e Sens. 2022,14, 1911 22 o 27 Acknowledgmen s: The au ho s would like o hank he Resea ch Suppo Founda ion o he Pe - nambuco S a e (FACEPE, B azil—APQ-0215-5.01/10 and FACEPE - APQ-1159-1.07/14), he Na ional Council o Scien i ic and Technological De elopmen (CNPq, B azil) and also unds h ough he ellowship o he Resea ch P oduc i i y P og am (CNPq 305286/2015-3, 304060/2016-0, 309681/2019- 7, and 303767/2020-0), and he Coo dina ion o he Imp o emen o Highe Educa ion Pe sonnel (CAPES, B azil - Finance Code 001) o he esea ch and s udy g an s. The au ho s a e also g a e ul o inancial suppo om he Po uguese Founda ion o Science and Technology (FCT), and he Uni e si y o Coimb a, Po ugal. In addi ion, we also would like o hank he anonymous e iewe s o hei insigh ul commen s, o which signi ican ly inc eased he alue o his s udy. Con lic s o In e es : The au ho s decla e no con lic o in e es . Abb e ia ions Summa y o all he symbols and ac onyms used in his pape . I em Desc ip ion aand bA e he calib a ion coe icien s CpSpeci ic hea o ai dWillmo ’s index o ag eemen DEM Digi al ele a ion model dT Nea -su ace ai empe a u e g adien eaAc ual a mosphe ic wa e apo p essu e esWa e apo sa u a ion p essu e ETaAc ual e apo anspi a ion GSoil hea lux GEE Google Ea h Engine HSensible hea lux HR Ins an aneous ela i e humidi y k on Ka man cons an LAI Lea a ea index LCCC Lin’s conco dance co ela ion coe icien LE La en hea lux LSE Land su ace emissi i y LST Land su ace empe a u e LULC Land use and land co e MAE Mean absolu e e o NDVI No malized Di e ence Vege a ion Index NSE Nash-Su cli e e iciency coe icien PBIAS Pe cen bias PCA P incipal componen analysis ah Nea -su ace ae odynamic esis ance o hea anspo RMSE Roo mean squa e e o RnNe adia ion Rn24 Daily ne adia ion R2Coe icien o de e mina ion SAVI Soil-Adjus ed Vege a ion Index T0Ins an aneous ai empe a u e TaAi empe a u e u* F ic ion eloci y u200 Wind speed a he heigh o 200 m VCVege a ion co e WP ecipi able wa e Remo e Sens. 2022,14, 1911 23 o 27 I em Desc ip ion z1and z2A e he wo heigh s be ween he su ace o he ancho pixels zom Momen um oughness leng h αsup Su ace albedo εaA mosphe ic emissi i y ΛE apo a i e ac ion λLa en hea o apo iza ion o wa e ρai Ai densi y ψmand ψhS abili y co ec ion ac o s o momen um and sensible hea , espec i ely Re e ences 1. A nan, X.; Leal, I.R.; Taba elli, M.; And ade, J.F.; Ba os, M.F.; Câma a, T.; Jamelli, D.; Knoechelmann, C.M.; Menezes, T.G.C.; Menezes, A.G.S.; e al. A amewo k o de i ing measu es o ch onic an h opogenic dis u bance: Su oga e, di ec , single and mul i-me ic indices in B azilian Caa inga. Ecol. Indic. 2018,94, 274–282. [C ossRe ] 2. Fe ei a, T.R.; Sil a, B.B.D.; De Mou a, M.S.B.; Ve hoe , A.; Nób ega, R.L.B. The use o emo e sensing o eliable es ima ion o ne adia ion and i s componen s: A case s udy o con as ing land co e s in an ag icul u al ho spo o he B azilian semia id egion. Ag ic. Fo . Me eo ol. 2020,291, 108052. [C ossRe ] 3. Mo o, M.F.; Nic Lughadha, E.; de A aújo, F.S.; Ma ins, F.R. A Phy ogeog aphical Me aanalysis o he Semia id Caa inga Domain in B azil. Bo . Re . 2016,82, 91–148. [C ossRe ] 4. Magalhães, K.D.N.; Gua niz, W.A.S.; Sá, K.M.; F ei e, A.B.; Mon ei o, M.P.; Nojosa, R.T.; Bieski, I.G.C.; Cus ódio, J.B.; Balogun, S.O.; Bandei a, M.A.M. Medicinal plan s o he Caa inga, no heas e n B azil: E hnopha macopeia (1980–1990) o he la e p o esso F ancisco Joséde Ab eu Ma os. J. E hnopha macol. 2019,237, 314–353. [C ossRe ] 5. de Medei os e Sil a, É.; Paixão, V.H.F.; To qua o, J.L.; Luna di, D.G.; de Oli ei a Luna di, V. F ui ing phenology and consump ion o zoocho ic ui s by wild e eb a es in a seasonally d y opical o es in he B azilian Caa inga. Ac a Oecologica 2020 ,105, 103553. [C ossRe ] 6. dos San os, L.R.; Nascimen o Lima, A.M.; Cunha, J.C.; Rod igues, M.S.; Ba os Soa es, E.M.; dos San os, L.P.A.; da Sil a, A.V.L.; Fe ei a Fon es, M.P. Does i iga ed mango cul i a ion al e o ganic ca bon s ocks unde agile soils in semia id clima e? Sci. Ho ic. 2019,255, 121–127. [C ossRe ] 7. de Souza Leão, P.C.; do Nascimen o, J.H.B.; de Mo aes, D.S.; de Souza, E.R. Yield componen s o he new seedless able g ape ‘BRS Ísis’ as a ec ed by he oo s ock unde semi-a id opical condi ions. Sci. Ho ic. 2020,263, 109114. [C ossRe ] 8. Rallo, G.; Paço, T.A.; Pa edes, P.; Puig-Si e a, À.; Massai, R.; P o enzano, G.; Pe ei a, L.S. Upda ed single and dual c op coe icien s o ee and ine ui c ops. Ag ic. Wa e Manag. 2021,250, 106645. [C ossRe ] 9. Gomes, L.S.; Maia, A.G.; de Medei os, J.D.F. Fuzzi ied hedging ules o a ese oi in he B azilian semia id egion. En i on. Chall. 2021,4, 100125. [C ossRe ] 10. Ma engo, J.A.; To es, R.R.; Al es, L.M. D ough in No heas B azil—Pas , p esen , and u u e. Theo . Appl. Clima ol. 2017 ,129, 1189–1200. [C ossRe ] 11. Ma ins, M.A.; Tomasella, J.; Rod iguez, D.A.; Al alá, R.C.S.; Gia olla, A.; Ga o olo, L.L.; Júnio , J.L.S.; Paolicchi, L.T.L.C.; Pin o, G.L.N. Imp o ing d ough managemen in he B azilian semia id h ough c op o ecas ing. Ag ic. Sys . 2018 ,160, 21–30. [C ossRe ] 12. de Quei oz, M.G.; da Sil a, T.G.F.; de Souza, C.A.A.; da Rosa Fe az Ja dim, A.M.; A aújo Júnio , G.D.N.; Souza, L.S.B.; Mou a, M.S.B. Composi ion o Caa inga Species unde An h opic Dis u bance and I s Co ela ion wi h Rain all Pa i ioning. Flo es a Ambien . 2021,28, 20190044. [C ossRe ] 13. Ba low, J.; Lennox, G.D.; Fe ei a, J.; Be engue , E.; Lees, A.C.; Nally, R.M.; Thomson, J.R.; de Ba os Fe az, S.F.; Louzada, J.; Oli ei a, V.H.F.; e al. An h opogenic dis u bance in opical o es s can double biodi e si y loss om de o es a ion. Na u e 2016 , 535, 144–147. [C ossRe ] 14. da Sil a Junio , C.A.; Teodo o, P.E.; Delgado, R.C.; Teodo o, L.P.R.; Lima, M.; de And éa Pan aleão, A.; Baio, F.H.R.; De Aze edo, G.B.; de Oli ei a Sousa Aze edo, G.T.; Cap is o-Sil a, G.F.; e al. Pe sis en i e oci in all biomes unde mine he Pa is Ag eemen in B azil. Sci. Rep. 2020,10, 16246. [C ossRe ] 15. da Rosa Fe az Ja dim, A.M.; da Sil a, T.G.F.; de Souza, L.S.B.; do Nascimen o A aújo Júnio , G.; Al es, H.K.M.N.; de SáSouza, M.; de A aújo, G.G.L.; de Mou a, M.S.B. In e c opping o age cac us and so ghum in a semi-a id en i onmen imp o es biological e iciency and compe i i e abili y h ough in e speci ic complemen a i y. J. A id En i on. 2021,188, 104464. [C ossRe ] 16. Cos a, M.D.S.; De Oli ei a-Júnio , J.F.; Dos San os, P.J.; Co eia Filho, W.L.F.; De Gois, G.; Blanco, C.J.C.; Teodo o, P.E.; da Sil a, C.A., J .; San iago, D.D.B.; Souza, E.D.O.; e al. Rain all ex emes and d ough in No heas B azil and i s ela ionship wi h El Niño–Sou he n Oscilla ion. In . J. Clima ol. 2021,41, E2111–E2135. [C ossRe ] 17. Tomasella, J.; Viei a, R.M.; Ba bosa, A.A.; Rod iguez, D.A.; San ana, M.D.O.; Ses ini, M.F. Dese i ica ion ends in he No heas o B azil o e he pe iod 2000–2016. In . J. Appl. Ea h Obs. Geoin . 2018,73, 197–206. [C ossRe ] 18. Viei a, R.M.S.P.; Tomasella, J.; Al alá, R.C.S.; Ses ini, M.F.; A onso, A.G.; Rod iguez, D.A.; Ba bosa, A.A.; Cunha, A.P.M.A.; Valles, G.F.; C epani, E.; e al. Iden i ying a eas suscep ible o dese i ica ion in he B azilian no heas . Solid Ea h 2015 ,6, 347–360. [C ossRe ] Remo e Sens. 2022,14, 1911 24 o 27 19. Ribei o, K.; de Sousa-Ne o, E.R.; de Ca alho, J.A.; Lima, J.R.D.S.; Menezes, R.; Dua e-Ne o, P.J.; Gue a, G.D.S.; Ome o, J.P.H.B. Land co e changes and g eenhouse gas emissions in wo di e en soil co e s in he B azilian Caa inga. Sci. To al En i on. 2016 , 571, 1048–1057. [C ossRe ] 20. de Cas o Teixei a, A.H.; Lei as, J.F.; And ade, R.G.; He nandez, F.B.T. Wa e p oduc i i y assessmen s wi h Landsa 8 images in he Nilo Coelho i iga ion scheme. IRRIGA 2015,1, 1–10. [C ossRe ] 21. Ronquim, C.C.; Lei as, J.F.; de Cas o Teixei a, A.H.; Sil a, G.B.; Ga çon, E.A.M. Wa e indica o s based on SPOT 6 sa elli e images in i iga ed a ea a he Pa aca u Ri e Basin, B azil. In Remo e Sensing o Ag icul u e, Ecosys ems, and Hyd ology XIX, 104211I; In e na ional Socie y o Op ics and Pho onics: Bellingham, WA, USA, 2017; Volume 10421, p. 104211I. 22. Cunha, J.; Nób ega, R.L.B.; Ru ino, I.; E asmi, S.; Gal ão, C.; Valen e, F. Su ace albedo as a p oxy o land-co e clea ing in seasonally d y o es s: E idence om he B azilian Caa inga. Remo e Sens. En i on. 2020,238, 111250. [C ossRe ] 23. Ba bosa, H.A.; Lakshmi Kuma , T.V.; Pa edes, F.; Ellio , S.; Ayuga, J.G. Assessmen o Caa inga esponse o d ough using Me eosa -SEVIRI No malized Di e ence Vege a ion Index (2008–2016). ISPRS J. Pho og amm. Remo e Sens. 2019 ,148, 235–252. [C ossRe ] 24. de Quei oz, M.G.; da Sil a, T.G.F.; Zolnie , S.; da Rosa Fe az Ja dim, A.M.; de Souza, C.A.A.; do Nascimen o A aújo Júnio , G.; de Mo ais, J.E.F.; de Souza, L.S.B. Spa ial and empo al dynamics o soil mois u e o su aces wi h a change in land use in he semi-a id egion o B azil. Ca ena 2020,188, 104457. [C ossRe ] 25. Fishe , J.B.; Mel on, F.; Middle on, E.; Hain, C.; Ande son, M.; Allen, R.; McCabe, M.F.; Hook, S.; Baldocchi, D.; Townsend, P.A.; e al. The u u e o e apo anspi a ion: Global equi emen s o ecosys em unc ioning, ca bon and clima e eedbacks, ag icul u al managemen , and wa e esou ces. Wa e Resou . Res. 2017,53, 2618–2626. [C ossRe ] 26. Souza, R.; Ha zell, S.; Feng, X.; An onino, A.C.D.; de Souza, E.S.; Menezes, R.S.C.; Po po a o, A. Op imal managemen o ca le g azing in a seasonally d y opical o es ecosys em unde ain all luc ua ions. J. Hyd ol. 2020,588, 125102. [C ossRe ] 27. Fend ich, A.N.; Ba e o, A.; de Fa ia, V.G.; de Bas iani, F.; Tenneson, K.; Guedes Pin o, L.F.; Spa o ek, G. Disclosing con as ing scena ios o u u e land co e in B azil: Resul s om a high- esolu ion spa io empo al model. Sci. To al En i on. 2020 ,742, 140477. [C ossRe ] 28. Lopes, V.C.; Pa en e, L.L.; Baumann, L.R.F.; Mizia a, F.; Fe ei a, L.G. Land-use dynamics in a B azilian ag icul u al on ie egion, 1985–2017. Land Use Policy 2020,97, 104740. [C ossRe ] 29. Blondeel, H.; Landuy , D.; Vangansbeke, P.; De F enne, P.; Ve heyen, K.; Pe ing, M.P. The need o an unde s o y decision suppo sys em o empe a e deciduous o es managemen . Fo . Ecol. Manag. 2021,480, 118634. [C ossRe ] 30. de A aujo, H.F.; Machado, C.C.; Pa eyn, F.G.; Nascimen o, N.F.D.; A aújo, L.D.; de A. P. Bo ges, L.A.; San os, B.A.; Bei igo, R.M.; Vasconcellos, A.; Dias, B.D.O.; e al. A sus ainable ag icul u al landscape model o opical d ylands. Land Use Policy 2021 ,100, 104913. [C ossRe ] 31. Liu, S.; Su, H.; Zhang, R.; Tian, J.; Chen, S.; Wang, W. Regional Es ima ion o Remo ely Sensed E apo anspi a ion Using he Su ace Ene gy Balance-Ad ec ion (SEB-A) Me hod. Remo e Sens. 2016,8, 644. [C ossRe ] 32. Mu i, P.R.; da Sil a, L.L.; Medei os, S.D.S.; Dub euil, V.; Mendes, K.R.; Ma ques, T.V.; Lúcio, P.S.; e Sil a, C.M.S.; Beze a, B.G. Basin scale ain all-e apo anspi a ion dynamics in a opical semia id en i onmen du ing d y and we yea s. In . J. Appl. Ea h Obs. Geoin . 2019,75, 29–43. [C ossRe ] 33. Teixei a, A.D.C.; de Mi anda, F.; Lei as, J.; Pacheco, E.; Ga çon, E. Wa e p oduc i i y assessmen s o dwa coconu by using Landsa 8 images and ag ome eo ological da a. ISPRS J. Pho og amm. Remo e Sens. 2019,155, 150–158. [C ossRe ] 34. Mo ei a, E.B.M.; Nób ega, R.S.; Da Sil a, B.B.; Ribei o, E.P. Es ima i a da e apo anspi ação em á ea u bana a a és de imagens digi ais TM-Landsa 5. Geosul 2019,34, 559–585. [C ossRe ] 35. da Sil a, M.V.; Pando i, H.; de Almeida, G.L.P.; de Lima, R.P.; dos San os, A.; da Rosa Fe az Ja dim, A.M.; Rolim, M.M.; da Sil a, J.L.B.; Ba is a, P.H.D.; da Sil a, R.A.B.; e al. Spa io- empo al moni o ing o soil and plan indica o s unde o age cac us cul i a ion by geop ocessing in B azilian semi-a id egion. J. S. Am. Ea h Sci. 2021,107, 103155. [C ossRe ] 36. Mhawej, M.; Faou , G. Open-sou ce Google Ea h Engine 30-m e apo anspi a ion a es e ie al: The SEBALIGEE sys em. En i on. Model. So w. 2020,133, 104845. [C ossRe ] 37. Júnio , J.B.C.; e Sil a, C.M.S.; De Almeida, H.A.; Beze a, B.; Spy ides, M.H.C. De ec ing linea end o e e ence e apo anspi a- ion in i iga ed a ming a eas in B azil’s semia id egion. Theo . Appl. Clima ol. 2019,138, 215–225. [C ossRe ] 38. IBGE Ins i u o B asilei o de Geog a ia e Es a ís ica. A ailable online: h ps://cidades.ibge.go .b /b asil/pe/pe olina/pano ama (accessed on 29 Augus 2021). 39. NASA Gio anni. Na ional Ae onau ics and Space Adminis a ion. A ailable online: h ps://gio anni.gs c.nasa.go /gio anni/ (accessed on 29 Augus 2021). 40. San os, C.; da Sil a, R.M.; Sil a, A.M.; Ne o, R.M.B. Es ima ion o e apo anspi a ion o di e en land co e s in a B azilian semi-a id egion: A case s udy o he B ígida Ri e basin, B azil. J. S. Am. Ea h Sci. 2017,74, 54–66. [C ossRe ] 41. Tasumi, M. P og ess in Ope a ional Es ima ion o Regional E apo anspi a ion Using Sa elli e Image y; Uni e si y o Idaho: Moscow, ID, USA, 2003. 42. Mole o-Lobos, I.; Ma a , C.; Ba ichi ich, J. Pe o mance o Sa elli e-Based E apo anspi a ion Models in Tempe a e Pas u es o Sou he n Chile. Wa e 2020,12, 3587. [C ossRe ] 43. Consoli, S.; Inglese, P.; Inglese, G. De e mina ion o e apo anspi a ion and c op coe icien o cac us pea (Opun ia icus-indica Mill.) wi h an ene gy balance echnique. Ac a Ho ic. 2013,995, 117–124. [C ossRe ] Remo e Sens. 2022,14, 1911 25 o 27 44. Liu, J.; You, Y.; Li, J.; Si ch, S.; Gu, X.; Nabel, J.E.M.S.; Lomba dozzi, D.; Luo, M.; Feng, X.; A ne h, A.; e al. Response o global land e apo anspi a ion o clima e change, ele a ed CO 2 , and land use change. Ag ic. Fo . Me eo ol. 2021 ,311, 108663. [C ossRe ] 45. Ha zell, S.; Ba le , M.S.; Po po a o, A. Uni ied ep esen a ion o he C3, C4, and CAM pho osyn he ic pa hways wi h he Pho o3 model. Ecol. Model. 2018,384, 173–187. [C ossRe ] 46. Laipel , L.; Ruho , A.L.; Fleischmann, A.; Kayse , R.H.B.; Kich, E.D.M.; Da Rocha, H.R.; Neale, C.M.U. Assessmen o an Au oma ed Calib a ion o he SEBAL Algo i hm o Es ima e D y-Season Su ace-Ene gy Pa i ioning in a Fo es –Sa anna T ansi ion in B azil. Remo e Sens. 2020,12, 1108. [C ossRe ] 47. Allen, R.; Wa e s, R.; Bas iaanssen, W.; Tasumi, M.; T ezza, R. SEBAL (Su ace Ene gy Balance Algo i hms o Land)—Idaho Implemen a ion, Ad anced T aining and Use s Manual, Ve sion 1.0; Idaho Depa men o Wa e Resou ces: Boise, ID, USA, 2002. 48. Bas iaanssen, W.G.M. SEBAL-based sensible and la en hea luxes in he i iga ed Gediz Basin, Tu key. J. Hyd ol. 2000 ,229, 87–100. [C ossRe ] 49. Bas iaanssen, W.G.M.; Noo dman, E.J.M.; Pelg um, H.; Da ids, G.; Tho eson, B.P.; Allen, R.G. SEBAL Model wi h Remo ely Sensed Da a o Imp o e Wa e -Resou ces Managemen unde Ac ual Field Condi ions. J. I ig. D ain. Eng. 2005 ,131, 85–93. [C ossRe ] 50. Allen, R.G.; Tasumi, M.; T ezza, R. Sa elli e-Based Ene gy Balance o Mapping E apo anspi a ion wi h In e nalized Calib a ion (METRIC)—Model. J. I ig. D ain. Eng. 2007,133, 380–394. [C ossRe ] 51. Cheng, M.; Jiao, X.; Li, B.; Yu, X.; Shao, M.; Jin, X. Long ime se ies o daily e apo anspi a ion in China based on he SEBAL model and mul isou ce images and alida ion. Ea h Sys . Sci. Da a 2021,13, 3995–4017. [C ossRe ] 52. Filho, W.L.F.C.; San iago, D.D.B.; de Oli ei a-Júnio , J.F.; Junio , C.A.D.S. Impac o u ban decadal ad ance on land use and land co e and su ace empe a u e in he ci y o Maceió, B azil. Land Use Policy 2019,87, 104026. [C ossRe ] 53. Rouse, J.W.; Haas, R.H.; Schell, J.A.; Dee ing, D.W.; Ha lan, J.C. Moni o ing he Ve nal Ad ancemen and Re og ada ion (G eenwa e E ec ) o Na u al Vege a ion. NASA/GSFCT Type III Final Repo ; NASA/GSFCT: G eenbel , MD, USA, 1974; pp. 1–390. 54. Gao, Q.; Li, Y.; Wan, Y.; Lin, E.; Xiong, W.; Jiangcun, W.; Wang, B.; Li, W. G assland deg ada ion in No he n Tibe based on emo e sensing da a. J. Geog . Sci. 2006,16, 165–173. [C ossRe ] 55. de Lima, I.P.; Jo ge, R.G.; de Lima, J.L.M.P. Remo e Sensing Moni o ing o Rice Fields: Towa ds Assessing Wa e Sa ing I iga ion Managemen P ac ices. F on . Remo e Sens. 2021,2, 762093. [C ossRe ] 56. Teixei a, A.H.C.; Bas iaanssen, W.G.M.; Ahmad, M.D.; Bos, M.G. Re iewing SEBAL inpu pa ame e s o assessing e apo anspi- a ion and wa e p oduc i i y o he Low-Middle São F ancisco Ri e basin, B azil: Pa B: Applica ion o he egional scale. Ag ic. Fo . Me eo ol. 2009,149, 477–490. [C ossRe ] 57. B igh , R.M.; Da in, E.; O’Hallo an, T.; Pong a z, J.; Zhao, K.; Cesca i, A. Local empe a u e esponse o land co e and managemen change d i en by non- adia i e p ocesses. Na . Clim. Chang. 2017,7, 296–302. [C ossRe ] 58. Bas iaanssen, W.G.M.; Pelg um, H.; Soppe, R.W.O.; Tho eson, B.P.; Allen, R.G.; Teixei a, A.H.C. The mal-in a ed echnology o local and egional scale i iga ion analyses in ho icul u al sys ems. Ac a Ho ic. 2008,792, 33–46. [C ossRe ] 59. Teixei a, A.H.C.; Bas iaanssen, W.G.M.; Mou a, M.S.B.; Soa es, J.M.; Ahmad, M.D.; Bos, M.G. Ene gy and wa e balance measu emen s o wa e p oduc i i y analysis in i iga ed mango ees, No heas B azil. Ag ic. Fo . Me eo ol. 2008 ,148, 1524–1537. [C ossRe ] 60. Filho, W.L.F.C.; De Oli ei a-Júnio , J.F.; De Ba os San iago, D.; De Bodas Te assi, P.M.; Teodo o, P.E.; De Gois, G.; Blanco, C.J.C.; De Almeida Souza, P.H.; da Sil a Cos a, M.; Gomes, H.B.; e al. Rain all a iabili y in he B azilian no heas biomes and hei in e ac ions wi h me eo ological sys ems and ENSO ia CHELSA p oduc . Big Ea h Da a 2019,3, 315–337. [C ossRe ] 61. Teixei a, A.D.C.; Bas iaanssen, W.; Ahmad, M.-U.; Bos, M. Re iewing SEBAL inpu pa ame e s o assessing e apo anspi a ion and wa e p oduc i i y o he Low-Middle São F ancisco Ri e basin, B azil: Pa A: Calib a ion and alida ion. Ag ic. Fo . Me eo ol. 2009,149, 462–476. [C ossRe ] 62. Al a es, C.A.; S ape, J.L.; Sen elhas, P.C.; de Mo aes Gonçal es, J.L.; Spa o ek, G. Köppen’s clima e classi ica ion map o B azil. Me eo ol. Z. 2013,22, 711–728. [C ossRe ] 63. Beck, H.E.; Zimme mann, N.E.; McVica , T.R.; Ve gopolan, N.; Be g, A.; Wood, E.F. P esen and u u e Köppen-Geige clima e classi ica ion maps a 1-km esolu ion. Sci. Da a 2018,5, 180214. [C ossRe ] 64. Oli ei a, P.T.; e Sil a, C.M.S.; Lima, K.C. Clima ology and end analysis o ex eme p ecipi a ion in sub egions o No heas B azil. Theo . Appl. Clima ol. 2017,130, 77–90. [C ossRe ] 65. da Rosa Fe az Ja dim, A.M.; da Sil a, M.V.; Sil a, A.R.; dos San os, A.; Pando i, H.; de Oli ei a-Júnio , J.F.; de Lima, J.L.; de Souza, L.S.B.; do Nascimen o A aújo Júnio , G.; Lopes, P.M.O.; e al. Spa io empo al clima ic analysis in Pe nambuco S a e, No heas B azil. J. A mos. Sola -Te . Phys. 2021,223, 105733. [C ossRe ] 66. P es on, W.; Nascimen o, C.; Sil a, Y.; Sil a, D.J.; Fe ei a, H.A. Soil e ili y changes in ineya ds o a semia id egion in B azil. J. Soil Sci. Plan Nu . 2017,17, 672–685. [C ossRe ] 67. Menezes, K.M.S.; Sil a, D.K.A.; Gou eia, G.V.; da Cos a, M.M.; Quei oz, M.A.A.; Yano-Melo, A.M. Shading and in e c opping wi h bu elg ass pas u e a ec soil biological p ope ies in he B azilian semi-a id egion. Ca ena 2019,175, 236–250. [C ossRe ] 68. Giongo, V.; Coleman, K.; da Sil a San ana, M.; Sal iano, A.M.; Olsz eski, N.; Sil a, D.J.; Cunha, T.J.F.; Pa en e, A.; Whi mo e, A.P.; Rich e , G.M. Op imizing mul i unc ional ag oecosys ems in i iga ed d yland ag icul u e o es o e soil ca bon—Expe imen s and modelling. Sci. To al En i on. 2020,725, 138072. [C ossRe ] 69. INMET Ins i u o Nacional de Me eo ologia. A ailable online: h ps://po al.inme .go .b / (accessed on 29 Augus 2021).