Jos′e Miguel Reichert,Paulo Ivonir Gubiani,Danilo Rheinheimer dos Santos,Dalvan Jose′ Reinert,Celso Aita,Sandro Jos′e Giacomini
Soils Department,Universidade Federal de Santa Maria(UFSM),Santa Maria,RS,Brazil
Keywords:Natural resources management Environmental monitoring Soil conservation Soil quality Alleviating poverty
A B S T R A C T Environmental monitoring of small,rural watersheds was one of the components of the Natural Resources Management and Rural Poverty Alleviation Program(RS-Rural)in southern Brazil.The purpose of the monitoring was to assess the impact of promoting soil conservation and environment management practices adopted by farmers and funded by the Program.In four small monitored watersheds,in a total of 95 plots representing distinct land use and soil management,surface soil was collected to characterize ground-zero of the Program by determining several soil physical,chemical and microbiological properties.Principal component analysis(PCA)shows soil physical,chemical and biological properties were decisive in defining the agricultural soils in the rural watersheds with family farming.The sensitivity to chemical properties provides an opportunity to improve soil quality if soil management focuses on altering those properties.Soil management practiced by tobacco farmers leads to rapid,intense degradation of some natural soil properties,especially those related to the dynamics of soil organic matter,compared with more conservationist uses(forest,regrowth,and grassland).Thus,soil management must be reoriented to avoid the progress of degradation and recover soil physical and biological quality.Cover crops and by land-abandonment to allow natural vegetation are important management strategies for the degraded soils used for tobacco production,increasing soil organic matter,nutrients and microbial activity and thus allowing further crop production.In conclusion,watersheds with tobacco cropping have soils with lower quality than when under no-tillage grain production,requiring changes in land use and soil management.
Family farming,i.e.,farms operated mainly by family members,in southern Brazil has important common traits.The farm units,through inheritance sharing,gave rise to smallholdings;practice of polyculture,such as cultivation of various crops and animal raising,destined to supply the family and surplus for the commercialization;make intensive use of natural resources,e.g.,crop production relying on soil natural fertility;and labor is provided by familymembers(Brum,1987,p.200).This author points out,among the factors that have most influenced the decay of small family farming,the exhaustion of soil fertility stands out,caused mainly by intense land use and absence of soil conservation and proper management;substantial reduction in rural properties size and in area useful for cultivation;and low prices of agricultural products,constantly debased in the agricultural market.
Environments such as small rural watersheds occupied by family farming are,in most cases,quite diversified(naturally or through anthropic use),which leads to greater variability in soil properties.In a given watershed,there are different soil classes,where each of them may be under different uses such as forests(native or exotic),grazing areas,transition areas(forest regrowth),perennial crops such as fruit,and areas cultivated with or without soil tillage.This natural and anthropic diversity produces a great variation in soil biological,physical and chemical properties.
To reduce poverty in rural areas,degradation of natural resources,and rural population exodus in the Rio Grande do Sul state,southern Brazil(Grando,2007),the Natural Resources Management and Rural Poverty Alleviation Program(RS-Rural),funded by the International Bank for Reconstruction and Development(IBRD)and RS State government,was executed from 1997 to 2005.The state had about 400,000 rural families,of which 90% are small farmers,including 234,000 families considered poor(Waquil &Mattos,2003).The RS-Rural Program sought to improve the quality of life and farm productive capacity,and to promote integrated actions of family and community infrastructure,income generation and management and conservation of natural resources.
The State Foundation for Agricultural Research was responsible for the impacts on natural resources,which outsourced this activity to the Federal Universities of Santa Maria and of Rio Grande do Sul.These two institutions monitored the evolution in soil and surface water quality in four small watersheds,the majority of which farmers benefited from RS-Rural resources.Watershed programs aimed at improving the use of natural resources and alleviating poverty must be evaluated for their concrete effects and benefits to farmers and society(Mengistu&Assefa,2019;Sriyana et al.,2020;Teka et al.,2020).
In these regions occupied by small family farms,the analysis of soil quality and sustainability is a challenging task,requiring the integration of numerous indicators to contemplate effects of various edaphoclimatic,agri-environmental,and socio-economic components.The soil,its uses and management practices,cultivated plants and agronomic technologies are distinct and diverse,and imply changes of different magnitude in the properties when assessing soil quality.Soil quality emerges from the operational level of biological,chemical and physical processes that are jointly responsible for the organization and functionality of the soil(Bünemann et al.,2018;van Es & Karlen,2019).Because of the difficulty in jointly evaluating processes such as water and gas flow,cycling and transport of nutrients and dynamics of biological species and organic matter,soil quality is evaluated through measures of its properties that indicate its organization and functionality for the respective processes(Reichert et al.,2003).
Differences in soil properties are largely defined by factors that governs pedogenesis,such as parent material,relief,vegetation,climate,and time.However,land use and management usually modify properties of different soil types,from structurally-fragile sandy(Vaz et al.,2005;Vogelmann et al.,2010;Fasinmirin &Reichert,2011;Reichert et al.,2021a)to clayey soils(Secco et al.,2009;Holthusen,Brandt,Reichert,& Horn,2018,b;Reichert et al.,2021b,c),and from lowlands(Goulart et al.,2021)to sloppy lands(França et al.,2021).Substituting cultivated species for native vegetation produces great variation in soil biological,physical and chemical properties(Reichert et al.,2014,2016,2017).
Various soil properties are closely related to organic matter and soil aggregation(Haruna et al.,2020;Pulido-Moncada et al.,2018;Tisdall & Oades,2010),mainly regarding macroaggregates,a dynamic fraction that requires a continuous supply of organic material to remain stable over time(Tisdall & Oades,2010).Soil structural stability undergoes temporary changes,due to cyclical variations caused by soil tillage practices and agricultural machinery traffic,climate,plant growth,and soil and crop management(Haruna et al.,2020;Wohlenberg et al.,2004).
Establishing the effect of management on soil structure is important to assess the adequacy of cultivation practices to the type of soil and environment.In homogeneous and controlled experimental areas,the effect of management practices is easily detected;however,when several sites with different soils and different management practices are evaluated together,the relationship between variables may lead to misinterpretation of the system.However,for such of great complexity affecting the relationships between variables,multivariate analysis can help explain the general relationship of a set of variables,and better describe the trends and pattern recognition of the system or environment under analysis(Manly,2008).Thus,integrative assessment of soil quality and proposition of better use and management of natural resources are possible.
The objective of this study was to assess soil physical,chemical and biological properties,and to evaluate agricultural soil quality in four small,rural watersheds under family farming in southern Brazil,either used mainly for tobacco or for no-tillage grain production.The results contribute to a better planning and management of natural resources by small farmers,particularly in areas under conventional-tillage tobacco production by promoting soil conservation and environment management practices by farmers in extension programs.The novelty of this study is the focus on soils in watersheds with small farming units in sloppy lands,which are abundant in agriculture worldwide,but are highly understudied and usually receive little attention by research,rural extension and development policy.In particular,the monitored watersheds are located in one of the most biodiverse biomes in the world,and herein we show the it is possible to integrate small-scale production with environmental preservation,since farmers historically distribute the plots cultivated discontinuously in the natural landscape(discontinuity of the plots,with maintenance of the natural biome or areas of natural revegetation).Nonetheless,there is still an urgent need to adapt conservationist management practices to the typical cropping systems in impoverished family farming.
2.1.Characterization of the studied watersheds
The data used in this study come from Ground-zero in 2001 and Mid-term reports in 2004 of the RS-Rural Program.Soil samplings were made in four small,rural watersheds in the state of Rio Grande do Sul,Brazil,monitored to assess the agri-environmental impacts of the application of financing resources.The results were presented to the two funding agencies and,after a grace period/exclusivity of the data,we now present them to the scientific community.
The studied watersheds were:
(a)Arroio Lino,located in the municipality of Agudo,with 480 ha occupied by 42 families.The predominant relief is strongly undulating and the soils are mostly shallow,including Cambissols,Entisols,and Ultisols,derived from basalt mixed with sandstone.
(b)Lajeado Ferreira-C^andido Brum,located in the municipality of Arvorezinha,with an area of 2500 ha occupied by 170 families.The relief is undulating and the soils are shallow,with predominance of Cambissols,Entisols,and Ultisols derived from basalt/rhyolite and riodacite.
(c)Passo do Meio,located in the municipality of Cristal,with an area of 890 ha occupied by 58 families.The relief is undulating,with more sandy soils classified as Cambissols and Entisols,derived from granite.
(d)Rio Inhandava,located in the municipality of Maximiliano de Almeida,has an area of 1196 ha occupied by 47 families.The relief is gently undulating,and the soils are more developed than those in the first three watersheds,consisting of Oxisols,Alfisols,and some Entisols,derived from basalt.
In Agudo,climate is humid subtropical without drought,Cfa;rainfall is usually well distributed,varying from 1300 to 1800 mm year-1;average temperature of the hottest month exceeding 22°C,and in the coldest month varying between-3 and 18°C.In Arvorezinha,climate is classified as Cfb(sub-tropical super-humid with no dry season and warm summer)according to the K¨oppen classification,and average annual rainfall is 1938 mm(15 years,2002/2016;Ramon,2017).In Crystal,climate is humid subtropical,without drought,of the K¨oppen Cfa type;average annual precipitation is 1.234 mm;average annual temperature of 18.9°C,maximum exceeds 30°C in some days during the summer,and low temperatures can reach below 0°C in some winter days;the hottest months are January and February,and the coldest are June and July.In Maximiliano de Almeida,climate is humid subtropical climate,of the K¨oppen Cfa type;rainfall(annual total of 1,680 mm)well distributed in all months of the year;annual average temperature is 18.7°C,with the average maximum temperature being 25.7°C and the average minimum temperature 13.2°C.
In the first three watersheds,spontaneous vegetation quickly establishes itself in areas no longer cultivated,recovering organic matter and maintaining higher levels of nutrient availability than in native forest(Stürmer et al.,2011).The predominant cash crop is tobacco(Nicotiana tabacumL.).The soil is conventionally tilled,with two plowings with animal traction,where the first is done approximately 30 days before planting and the second precedes ridging.Constraints arising from this agriculture model include soil erosion,degradation of natural resources,and productivity resulting from soil degradation(Reichert,Pellegrini,& Rodrigues,2019,b),and pollution of water supply sources and water bodies(Bastos et al.,2021;Becker et al.,2009).The technological package of tobacco companies is applied to all producers homogeneously without considering social aspects and soil properties.The vast majority of tobacco growers use lime at intervals of time compatible with technical recommendations.The fertilization system,however,disregards soil analysis and the history of soil fertilization.The fertilizers are delivered to farmers together with all the necessary inputs for the implantation,conduction,harvest and processing of the tobacco crop,and thus farmers have no power against the recommendation package.The fourth watershed has cultivated soils with high fertility,especially no-tillage and crop rotation.In addition to less soil and nutrient losses due to water erosion in conservation systems,the use of legume species in rotation systems allows nitrogen to be incorporated into the soil.Another factor to be considered is the frequent application of inputs by farmers,both industrialized fertilizers and the use of animal waste,which maintains and even raises the levels of nutrients in the soil.
The three watersheds under conventional tillage(tobacco),and one under no-tillage(grains).There are no major variations in soil management among producers in the basins,since impoverished family farmers use similar,conventional tillage,while those adopting no-till for grain production also use similar management practices.The only soil conservation practice for tobacco,at the time of collection,was the construction of ridges across the slope,acting as micro-terraces in the tobacco fields.
2.2.Soil sampling and analysis
Soil samplings,either disturbed soil or core samples,were carried out in the 0—5 cm layer,in a total of 95 plots representing the land uses and management of the four monitored watersheds.The plots are fragmented and interwoven in the landscape,either from the natural biome(Atlantic Forest)or from natural revegetation after deforestation.In the three tobacco-producing watersheds,plots are a maximum of 1 ha,while in the grain-producing watershed the plots are larger(about 5 ha).
The total number of samples is defined by a combination of four watersheds,with five land uses(cropland,grassland,native forest,planted forests,regrowth forest),resulting in 17 combinations.The sample sizes for each of these 17 combinations varies between 1 and 20.Thus,a balanced sampling design was not defined.Several locations were under sampled(one plot,Fig.1)because they were very small areas compared to the large areas under annual crops.Nonetheless,a balanced sampling design is not mandatory to apply the PCA analysis.
Sampling at Ground-zero in the three tobacco watersheds was carried out in August/September;therefore,the fields were fallow or with winter plants.In the Mid-term,sampling in Arvorezinha and Agudo was carried out in October,which corresponds to the beginning of the tobacco transplant,while in Maximiliano de Almeida sampling was done in December,when corn was tasseling and soybeans in stage V2—V4.
The following soil properties were determined:
1)sand,silt and clay,determined by the pipette method(Embrapa-Empresa Brasileira de Pesquisa Agropecu′aria,2007)after dispersion using the procedure of Suzuki et al.(2015),and water dispersible clay WDC(Embrapa-Empresa Brasileira de Pesquisa Agropecu′aria,2007);
2)soil bulk density,total porosity e microporosity on core samples(Embrapa-Empresa Brasileira de Pesquisa Agropecu′aria,2007);
3)saturated hydraulic conductivity on core samples,using a constant head permeameter(Embrapa-Empresa Brasileira de Pesquisa Agropecu′aria,2007);
4)mean weight diameter(MWD)of water-stable aggregates,following the method of Kemper and Chepil(1965);
5)total organic carbon(TOC),according Nelson and Sommers(1982);
6)microbial biomass carbon(MBC),following the method of Islam and Weil(1998);
7)basal respiration,using the static method of CO2measurement,without air injection(Jenkinson and Powlson,1976);
8)total nitrogen,pH in water,available P and K,and exchangeable Al,Ca and Mg,following procedures described in Tedesco et al.(1995);
9)soil covered with crop residues,crops,weeds and rocks/stones,total soil coverage and bare soil,all in percentage,using the method of Spedding and Large(1957);
10)potential acidity(H+Al),effective cation exchange capacity(CEC),and ratio Cbio/TOC,estimated from previous measurement.
2.3.Statistical analysis
The variables were initially assembled in a single data set(n=95)and summarized by descriptive statistics(mean,median,maximum,minimum,coefficient of variation,asymmetry and kurtosis).Pearson"s linear correlation analysis on the total data set was performed,and multiple regression was done for MWD of water-stable aggregates as a function of soil biological,physical and chemical properties,with the stepwise procedure(SAS Institute,1999).
Subsequently,the values were grouped by watershed and land use for the application of principal component analysis(PCA).The land uses associated with their respective watershed constituted the individuals,called areas for better correspondence with the object of study(Fig.1),whose variables were represented by the mean value.Thus,a data matrix X(nxp)of order 17×22 was organized,with the areas(n=17)arranged in the rows,and the variables(p=22)arranged in the columns.The data in matrix X(17x22)were normalized(z scores),so that each variable had a mean equal to zero and variance equal to one.Auto-scaling of data is justified since the variables have different scales(units of measure)and,therefore,differ in order of magnitude,which would cause too much influence of a few variables in the PCA model(Sena et al.,2002).
Fig.1.Watersheds and land uses studied.
The matrix of scaled,X(17x22),and scaled data,Z(17x22),were submitted to Pearson"s linear correlation analysis.Variables with three or more significant correlations were maintained for PCA,and the rest was removed from the set as they are of little importance for analysis when little correlated(Manly,2008).The final matrix of the scaled data Z(nxm)had dimension Z(17x16)(due to the exclusion of six variables).The PCA was performed on the matrix Z(17x16).
3.1.General statistics,correlation and regression analysis
The descriptive statistics of the total set of soil analyses results indicate many variables do not strictly follow a normal distribution,i.e.,asymmetry between-1 and 1 and kurtosis equal to 3(Table 1).However,the mean deviates substantially from the median only in cases where the asymmetry is less than-1.4 or greater than 1.4,while kurtosis(flattening of the normal distribution curve)describes with less precision the deviation between mean and median.Thus,parametric analyses can be inefficient for 43% of the variables,since the parameters(mean,standard deviation and variance)do not adequately represent the variable.
Table 1Descriptive statistics for soil physical,chemical and microbiological properties from the four watersheds.
Although the coefficient of variation(CV)is affected by the asymmetry of the distribution(for some variables),the data variability can still be assessed by the CV,but not so precisely.Thus,three groups of variables can be distinguished in terms of variability.Variables with less than 35% variation(MWD of waterstable aggregates,total soil coverage,pH and exchangeable K),variables with variation greater than 75%(soil covered with crop residues,bare soil,MBC,total N and exchangeable Al)and variables with variation between 35 and 75%(all others).
The results of Pearson"s linear correlation analysis on the total data set(n=95)were summarized to represent the number of significant correlations of each variable with the others,and the extreme correlation coefficients with the respective variables(Table 2).Similarly,the correlations of the auto-scaled variables are also shown in Table 2.Both the number and the degree of correlations increase from top to bottom,indicating that the base variables are better associated with the other variables.The number of significant correlations decreased for the scaled data,due to the reduction of the variance and sample size(n=17).
3.2.Multivariate analysis
After auto-scaling,there are 16 variables with three or more significant correlation(Table 2),which appear in the PCA correlation diagram(Fig.3).The other six variables shown in Table 2(total coverage,crop cover,residue cover,total N,MWD of stable aggregates,and microbial biomass C/TOC)were,therefore,not included in the PCA analysis.
Variables with the lowest correlation after scaling are practically the same as before scaling,with the exception of MWD of waterstable aggregates and MBC/TOC.On the other hand,multiple regression indicated that 50% of the MWD of water-stable aggregates variation was explained by seven variables(highest R2=0.499).Thus,limited information on the effects of soil management and use on the structural quality was obtained by regression analysis.
The first two main components(PC1 and PC2)of the Principal Component Analysis contain 64% of the data variance(the sum of the first two eigenvalues)and represent the general environment pattern(Table 3).With two more components(PC3,with 12%,and PC4,with 10% of the variance),86% of the data variance would be considered to explain the relationships between the areas.All the other 11 components explained altogether only 25%of the variance,and were considered of little importance to describe the main structure of the data.Since PC3 and PC4 contributed little to thedistinction of the areas,the diagrams of the first two components were presented(Fig.3)and only the main contrasts in PC3 and PC4 were commented.
Table 2Correlation among soil physical,chemical and microbiological properties from the four watersheds,before and after the scaling.
Table 3Variability descriptors and the pattern of relationship of areas and variables.
The scores for each area(watershed_use)for PC1 are obtained by adding the products of the coordinates of the first eigenvector by the respective variables,e.g.,score of area 1 of PC1=0.0304*Clay+0.1407*Silt+…-0.1589*Bare soil).The procedure is repeated for the calculation of the PC2 and PC3 scores.Thus,it appears that the eigenvector is a matrix of weights applied to the variables.The variables that receive greater weight have a greater contribution to area positioning in the orthogonal plane formed by the components.
3.2.Multivariate analysis for soil chemical,physical and microbiological fragility
Soil fertility properties are the basic variables contrasted in PC1(Table 3 and Fig.3a).The highest positive weights are for exchangeable variables K,Ca and Mg,CEC and basal respiration,while the most negative weights are for exchangeable variable Al(Table 3).Thus,areas in which low soil fertility is the predominant characteristic(higher values for exchangeable Al)are located to the left in relation to the origin of PC1,as their scores are negative(Fig.3b),whereas areas with good soil fertility and higher microbial basal respiration are positioned to the right of the PC1 origin.
Soil physical fragility and associated chemical properties are contrasted in PC2(Table 3 and Fig.3a).The highest positive weights indicating this characteristic are related to clay,silt,MBC and H+Al(potential acidity),while the most negative weights are related to sand and available P.Thus,areas in which physical fragility and low P availability are the predominant characteristics(higher values for sand)are positioned below the origin of PC2(Fig.3b),whilst areas with less physical fragility and high potential acidity and MBC are positioned above the origin of PC2.By contrast,TOC(eigenvector=-0.2914)and bare soil(eigenvector=0.5432)weighed most for contrasting areas in PC3(Table 3),while in PC4 no additional variable to those already mentioned in PC1,PC2 and PC3 was important to explain the contrast between areas(Table 3).
Basal respiration and MBC had greater variation than for chemical properties,although the partial variability is mainly linked to physical properties.When the self-vector is analyzed,the dimension is similar among the physical,chemical and biological properties(Table 3).Thus,compared with chemical and biological properties,it seems that physical properties better discriminate between different sites.Furthermore,in PC2,the MBC is positively correlated with clay,silt and WDC,demonstrating that soils with higher clay content also have higher MBC.
4.1.Soil properties relationships
The data set was derived from different pedoenviroments in which land uses were considered similar,although there were several soil management systems adopted by farmers in each watershed.In an uncontrolled environment,the Cartesian relationship between two variables is not always evident.For instance,soil aggregate stability herein represented by MWD of water-stable aggregates,a soil property used to differentiate management effects(Lopes et al.,2020),had low correlation(-0.22 to 0.52)with seven of the 22 variables,including soil properties and land cover(Table 2).This low dependence of MWD to the predictors in the regressions(Fig.2)resulted in its exclusion from the PCA.
The low coefficient of variation of MWD of water-stable aggregates(Table 1)already indicated a weaker covariance relationship between this property and the other analyzed properties.Soil MBC,soil coverage by crop,total soil coverage and the basal respiration were positively and significantly related to MWD of water-stable aggregates(Fig.2).This aspect highlights the importance of carbon supply for soil structure improvement,since soil aggregation is improved by mechanical binding of soil particles by microbial cells and hyphae and aggregating effect of products derived from microbial synthesis or from organic residues decomposition(Demenois et al.,2018;Lopes et al.,2021).
4.2.Comparing watersheds
Comparisons among the small watersheds and soil management can be done by PCA.The positioning of the watersheds in PC2 shows soils in Agudo and Cristal watershed are physically more fragile,and the fragility is more pronounced in areas with annual crops,i.e.,CRT_AC and AGD_AC,as these are at the extreme of PC2.The high environmental fragility is governed by particle size(low silt and clay,and high sand content),and their consequence with soil organic carbon dynamics(Fig.3a).
Fig.2.Stepwise procedure summary of the multiple regression analysis between mean weight diameter(MWD,mm)of water-stable aggregates and soil biological,physical and chemical properties.
Destruction of the natural biome(Atlantic Forest,in the case of our study)to transform into agroecosystems leads inexorably to drastic reduction in soil organic carbon stock(Stürmer et al.,2011).Use of conventional soil management systems for long periods,without adoption of conservation practices,increases entropy and physical and biological degradation,making the environment increasingly fragile.The variables TOC,MBC,and potential acidity are normally positively associated with soils with a higher clay content(Theng et al.,1989),while the positive relationship of water dispersible clay with total clay must be associated with the natural fragility of prevalent soils in the watersheds.Soils with higher clay content,but physically fragile,will have a large amount of water dispersible clay.Another aggravating factor is the excessive fertilization of fertilizers containing sodium(clay dispersing agent),such as sodium nitrate used for tobacco cultivation at high doses(Kaiser et al.,2010).
The positive correlation of available phosphorus with the condition of physical degradation results from two main factors.The first is related to high applications of phosphate fertilizers in the soil,being well above the critical agronomic level and contrasting with low levels in non-anthropized areas(Rheinheimer et al.,2008),especially in the areas of tobacco cultivation or with use of animal waste.The second is related to the Mehlich I method,which has greater extracting power for this nutrient in soils with low clay content(Novais & Smyth,1999).
In the Arvorezinha and Maximiliano watersheds,although the soils are less physically fragile,annual crops are also leading to greater soil fragility,as the areas with these uses(AVZ_AC and MXL_AC)have been displaced down on PC2.In the Agudo,Cristal and Arvorezinha watersheds,areas with annual crops are shifted to the left on PC2 in relation to areas with conservationist uses(forest,regrowth,and grassland),indicating that the soil management practiced by farmers leads to rapid,intense degradation of soil properties,especially those related to soil organic matter dynamics.For example,reforested areas in the Cristal watershed(CRT_FP)are present on highly-degraded soils from continuous cultivation with tobacco without any concern for maintaining the natural soil organic carbon stock.
In Arvorezinha,soil degradation caused by tobacco cultivation(AVZ_AC)is less evident than observed in the other two tobaccoproducing watersheds(Agudo and Cristal).This is most probably related to recent deforestation and the financial incentive provided by the RS-Rural Program for the adoption of soil conservation practices,even for tobacco cultivation(Didon′e et al.,2014;Minella et al,2007,2009).In contrast,in the Maximiliano Almeida watershed,annual cropping(MXL_AC)has even improved chemical fertility,even though the cultivated areas have a lower index of soil coverage compared with natural or regenerated environments.In this watershed,upland grain crops predominate under no-tillage system since the mid-90s of the last century.
Tobacco is cultivated almost exclusively in a conventional system with soil tillage before transplantation,without crop rotation and with intense use of fertilizers and pesticides.The amounts of added nutrients are,in most situations,much higher than crop requirements(Kaiser et al.,2010),following the philosophy of safety adopted by tobacco companies.As tobacco crop grows,several mechanical and chemical weedings are carried out,reducing soil coverage and protection against erosion.In two of the watersheds(Agudo and Arvorezinha)with tobacco monoculture,we observed high surface runoff and soil erosion(Bonum′a et al.,2014),both in crops and on roads(Minella et al.,2007),with contamination of water resources with nitrate(Kaiser et al.,2015;Bastos et al.,2021),phosphorus(Pellegrini et al.,2010;Tiecher et al.,2017),pesticides(Bortoluzzi et al.,2007;Becker et al.,2009;Sequinatto,Reichert,Rheinheimer,Reinert,& Copetti,2013),and biological agents such asEscherichia coli(Bastos et al.,2021).Notwithstanding,the adoption of conservationist management,soil and water losses are reduced in these watersheds(Minella et al.,2009).
In the watersheds occurring on fragile soils and with an absolute predominance of tobacco crop(Cristal,Agudo and Arvorezinha),based on the situation described by PCA,soil management must be reoriented to avoid the evolution of degradation,and recover soil physical and biological quality.Increasing carbon levels is one of the most effective ways to restoring the productivity of eroded soils(Oldfield et al.,2019).Although tobacco cultivation has severe impacts on soil properties through the imbrication of small cultivation areas in the landscape,the consequences on the environment are mitigated and even less impacting on watercourses than,for example,grain-producing crops even under no-till systems(Troian,2020).
Fig.3.Correlation between main components and physical,chemical and microbiological properties of soil sampled in areas with different land uses in four small watersheds(a),and area scores for the first and second components(b).In plot b,the watersheds are:AGD=Agudo,AVZ=Arvorezinha,CRT=Cristal,and MXL=Maximiliano de Almeida.Land uses are:AC=annual crops,GL=grassland,FN=native forest,FP=planted forest,and RG=natural forest regrowth.
For Cristal and Agudo waterseds,where soils contain more sand,the increase in aggregate stability improves soil physical quality,which is an important aspect especially Permanent crops or permanently covered soil are strategies that introduce organic material into the system in a more continuous way,enhancing the beneficial effect roots(Demenois et al.,2018)and products of microbial decomposition and metabolism to improve aggregation(Lehmann et al.,2017;Liu et al.,2019).Annual pastures,as far as possible,should be replaced by perennial pastures,preferably grasses with high density of roots.Periodic renewals of the root system and exudates in the soil stimulate microbial activity and provide greater formation and stabilization of aggregates(Liu et al.,2019).
Temporary use with tobacco cultivation with subsequent abandonment of the land for the recovery of natural vegetation is not a totally condemned practice.In fact,the recovery of natural vegetation is tremendously stimulated and,in less than a decade,soil organic C,N and S stocks equals those observed in natural pedoenvironments(Stürmer et al.,2011),even higher for other bioavailable stocks nutrients.In fact,the highest basal respiration was observed in natural forest regrowth areas in the Agudo watershed,which were accompanied by an increase in MBC.The re-cultivation of these“regenerated”lands allows further cropping of tobacco and other subsistence crops in the watersheds.Furthermore,slash-and-burn agroecosystems is even considered agroecologically adapted for poor regions(Kleiman et al.,1995),and to the maintenance of natural vegetation(Randriamalala et al.,2015).Afforestation in also an important practice to reduce soil erosion and sedimentation,and contribute to streamflow regulation in the watersheds(Rodrigues et al.,2018;Valente et al.,2020,2021).Finally,agriculture in South America either continues with conventional tilled in areas of multi-cropping family farming or evolves towards a“simplified”no-tillage grain production system,including crop rotation and physical barriers to control runoff.
Principal component analysis(PCA)show soil physical,chemical and biological properties were decisive in defining the agricultural soils in the rural watersheds with family farming.The sensitivity to chemical properties provides an opportunity to improve soil quality if soil management focuses on altering those properties.
Soil management practiced by tobacco farmers leads to rapid,intense degradation of some natural soil properties,especially those related to the dynamics of soil organic matter,compared with more conservationist uses(forest,regrowth,and grassland).Thus,soil management must be reoriented to avoid the progress of degradation and recover soil physical and biological quality.
Crop rotation,cover crops and by land-abandonment to allow natural vegetation are important management strategies for the degraded soils used for tobacco production,increasing soil organic matter,nutrients and microbial activity and thus allowing further crop production.
Acknowledgments
This study was financed by the International Bank of Reconstruction and Development(IBRD)through the RS-Rural Program,and the Brazilian Council for Scientific and Technological Development(CNPq)provided fellowships for the authors.
推荐访问:Land planning characterization