Water Quality Assessment for Organic Matter Load in Urban Rivers Considering Land Use Dynamics

doi:10.21203/rs.3.rs-2129806/v1

There has been an increasing number of global model studies that relate water quality to the watershed land use. The strategy of considering a model that is comparable to the Soil Conservation Service Curve-Number method that employs land use maps to estimate the effects of land use on the water quality, has considerable potential for application. In this context, this paper presents the LUPC (Land Use Pollutant Contribution) Model that relates pollutant loads originated by several types of coverage and a weight function which determines the load decay through water basin. It makes an estimate of the watershed features that are produced by conducting an analysis of the river profile, which is shown here in terms of the labile and refractory organic matter (OM), differentiated by its percentage of aromaticity. The initial results which were obtained by applying the model in a river basin located in the south of Brazil, suggest that the model can estimate the load of OM at a point of interest in the river and has proved to be a valuable tool for enabling managers to make decisions about the use planning land use and its impacts for water resources planning and management.

model

supporting decision-making tool

water resources management

Water is a natural resource that is essential for both human beings and maintaining a balanced environment. Controlling levels of pollution is becoming increasingly important when defining goals and taking measures for the future, since this is a valuable means for strategic water resources planning and management.

However, the continuous water quality monitoring actions can incur high financial costs (Li et al., 2017) and potentially can impair those agencies that have limit access to appropriate monitoring plans. Thus, the need for better ways for monitoring services considering secondary information available, such as Geographic Information Systems (GIS), can be useful for proposing rivers depollution programs basis improving monitoring plans.

In this context, the anthropic urban occupation of a water basin have both direct and indirect effects on river water quality, requiring models that can estimate the extent of this impact through satellite images (Giri & Qiu, 2016). The land cover reflection from the incident solar radiation on distinct electromagnetic spectrum bands can be analyzed, where it is possible to obtain information about the kinds of occupation or landscape that can be found among the forms of radiation among the wavelengths available (Asming et al., 2022; Trujillo-Jiménez et al., 2022).

On the other hand, considering the hydrological response of a watershed, the Curve Number Model from Soil Conservation Service (CN-SCS) is widely employed because of its robustness despite its simplicity (Kumar et al., 2021; Walega et al., 2020). The method allows estimating flows from rainfall input data. The model is based on the assumption that the permeability of a particular area depends on the soil type and the land use and land cover, that influences water storage that infiltrates and the fraction that allows superficial flow (Mishra & Singh, 1999)

On a comparable way, the land cover (urban, field, agricultural or forest) will influence the type and number of compounds, which will be conveyed to the water bodies of the watershed establishing a correlation between the type of spatial occupation and the quality of its water. Thus, establishing a model that is like the CN-SCS model and which envisages a means of estimating the pollutant load in water bodies can be applied because satellite images for classifying the types of soil cover will allow a considerable reduction in monitoring costs.

Additionally, the model framework can be adapted to incorporate depollution measures scenarios, integrated to potential changes from land use (i.e., deforestation regions for residential or agricultural areas) as a tool for water resources planning and management plans. Wang and Zhang (2018) highlights that this strategy can be considered a topic of critical research. Additionally, the proposed model for use with aquatic pollutants, should take account of the decay-accelerating factor of these substances, either through their degradation or deposition throughout the period between the origin of the pollutant and the point of interest in the water body. Among other factors, this decay is a factor in the distance between these two points.

Some already existing models can serve as a point of reference, such as the case of the Geographically Weighted Regression (GWR), described as a general kind of linear regression with the inclusion of a spatial component (Rimba et al., 2021; Wilson, 2015) or the model for the Inverse Distance Weighting (IDW) Interpolation (Bostanmaneshrad et al., 2018; Mir et al., 2017). Both mentioned models use a weight function to add less magnitude to a situation where there is a greater distance to the point of interest.

In the case of the GWR, the linear and angular coefficients incorporated the spatial variability of the data so that they could include in its formulation, a Kernel Function which associates the distance between the data region and the regression point is greater, assigns a lower weight to the coefficients. The Kernel function can take on a format of various distributions such as normal, exponential or bisquared (Koh et al., 2020; Tu & Xia, 2008; Zhang et al., 2021). Regarding the IDW, the parameters in a point of interest are obtained by the weighted average interpolation regarding both the distance of the surrounding data and the inverse of the distance (between the data point and prediction location) raised to a determined power such as that which the name itself describes (de Mesnard, 2013).

Thus the model proposed in this research will be based on the following assumptions: (i) each type of land cover produces a particular pollutant load (unit of mass by unit of time) by the unit of area (kg/(day m²)) or (ton/(day m²)); (ii) a determined fraction of the constituent load that is produced, leads to the total amount of that constituent in a particular point of interest in the river; (iii) the constituent fraction will be lower when the distance between the area under analysis and the point of interest in the river is greater; (iv) the constituent fraction can be determined by a weight function (Kernel Function), as in the GWR model, since the distribution functions, notably the Gaussian, tend to recur in natural phenomena; (v) the pollutant load in a determined point of a water body is the equivalent of the sum of all the fractions of this pollutant that emerge from a contributory basin. The model outlined here will hence forward be designated as the Land Use Pollutant Contribution (LUPC) and the specific pollutant load produced in each type of soil cover will be designated the Pollutant Contribution Index (PCI).

Different parameters for water quality can be employed to define the LUPC model, although the organic matter is present in a wide range of natural and anthropic sources, which are the aquatic biotic energy and involved in various chemical and biological processes. As well as this, it is one of the main indicators of pollution used around the world, as in the case of Brazil, and based on parameters of indirect determination, such as Biochemical Oxygen Demand (BOD) and Chemical Oxygen Demand (COD) or direct, such as Dissolved Organic Carbon (DOC) (Jouanneau et al., 2014). Thus, in this first phase, the LUPC model will be applied for the determination of the load of organic matter in rivers.

However, considering the great molecular complexity of organic matter, which is largely the result of its origin, there is a need to define how to quantify its fractions since they can be divided into labile component which have an easy degradation, while refractory fraction decomposition can take thousands of years. The BOD is widely used as a measure of biodegradable or labile organic matter; but its results show a wide variance, being permissible up to 20%, as well as the fact that it requires 5 days to obtain it. However, variations of up to 50% were obtained in the BOD values for real samples when these were analyzed in different laboratories (APHA, 2005; Kumar & Kumar, 2005).

On the other hand, the analysis of DOC allows a more precise quantification of the dissolved organic matter that was present in a sample, although this entailed a global analysis that was not divided into different fractions (Visco et al., 2005). Since each kind of land cover produces different proportions of organic matter labile and refractory fractions and this show differentiated features of degradation and deposition for the model on the screen, there is needed to ensure these fractions can be separated. However, there is no well-defined method for conducting this separation in a suitable way.

In this context, spectroscopic techniques of absorbance and fluorescence in the ultraviolet and visible range, are widely employed for characterizing and differentiating the fractions of organic matter, mainly because it is simple to apply them. These kinds of methods are generally used in a qualitative manner and reveal what the predominant fraction is in a sample (Ahmad & Reynolds, 1999; Carstea, 2012; Peuravuori et al., 2002; Rostan & Cellot, 1995; Westerhoff & Anning, 2000). At the same time, several studies have sought to make use of these techniques in a quantitative way because the larger the number of double or triple bonds in its molecule, the greater the absorbance of Ultra-Violet visible light, as well as the greater its fluorescence (Ahmad & Reynolds, 1999; Hudson et al., 2008; Hur & Kong, 2008; Edzwald et al., 1985).

Clearly, molecules of greater complexity like those with a high degree of aromaticity, have higher values of absorbance and fluorescence as is the case with humic substances, while simpler substances, such as proteins, that have predominantly aliphatic chains, absorb Ultra-Violet Visible light at a lower intensity (Carstea et al., 2012; Evans et al., 2005; Rostan & Cellot, 1995; Westerhoff & Anning, 2000; Wilkinson et al., 1997; Zumstein & Buffle, 1989).

Some spectroscopic indices are widely used for characterizing organic matter, as is the case of the specific ultraviolet absorptivity in 254 nm (SUVA₂₅₄) which makes use of the absorbance of light in this standardized wavelength through the concentration of carbon (DOC) to assess the predominant fraction of organic matter in a sample. SUVA₂₅₄ values of 4.4 were obtained for fulvic acid samples, which is a humic substance with the characteristics of refractory materials, while values close to 1.2 were obtained for samples of domestic sewage with predominantly labile features. Thus, the highest values of SUVA₂₅₄ show a predominance of refractory organic material, whereas the lowest values show a greater presence of labile organic matter (Her et al., 2008; Westerhoff & Anning, 2000).

Complementarily, the value of SUVA₂₅₄ can be used to indicate the degree of aromaticity of an organic matter sample (Weishaar et al., 2003). This means the UV-Visible absorption has the capacity to separate the labile and refractory fractions of organic matter, since it can be used in the quantification of its aromaticity degree, since this is one of the key factors that have an influence on its degradation.

On the basis of all that is outlined here, the main purpose of this study is to set out a new model for estimating the extent of pollutants in rivers, in particular the organic matter that is formed from the sum total of the loads produced by different soil covers, which are weighted by the measures obtained in terms of the distance between the centroids of the producing regions and the point of interest in the river. The secondary objectives include the following: i) the definition of the means of separating the labile and refractory fractions of organic material so that they can be used in the model; and ii) the determination of specific loads of labile and refractory organic matter produced by the types of soil cover in the analysis (PCI).

The CN-SCS model evaluates the rainfall fraction that contributes to the constituent dynamics up to the monitoring reference point, taking into account the total permeability of the basin when based on the fraction of each combined type of land cover found in it (Mishra & Singh, 1999). The proposed model is analogous to traditional schemes for assessing the pollution potential capacity of a water body based on the land cover of its water basin as summarized by Ferreira et al (2018), as consequence of each type of land cover (urban, forested, agricultural) that produces different loads of pollutants which are carried inside the main water body, as shown in Figure 1:

However, unlike the water, which only undergoes retention by the soil and vegetation, most of the pollutants also tend to be degraded along this journey and this reduction should be included in the load for the model. Thus, LUPC model includes the pollutant load caused by the unit of area in each kind of soil cover (j), represented by an index like the CN of the CN-SCS model, defined as PCI_j. Equation 1 presents the LUPC model:

Where: (i) The area occupied by this feature (A_j), is henceforward designated as the feature generation (j), together with a weight (w_ij) that which will determine what fraction of the pollutant load generated in the j feature, will be conveyed to the water course. The total load (Load_Pi) at the point of interest (Pi) will be equivalent to the total sum of the fractions of pollutants carried by the feature generators located in its standard basin; (ii) the weight w_ij will be inversely proportional to the distance (d_ij) between the centroid of the feature generation j and the point of interest Pi in the river and can be obtained by a Kernel Function, as in the case of the GWR model, which is represented in this study by a Gaussian Function (Equation 2), since many natural phenomena are well represented by the normal function (Huang et al., 2015; Rimba et al., 2021; Zhao et al., 2015).

The bandwidth b defines in what way the decay of the Kernel Function weights will take place - the greater the b value and the greater the bandwidth of the function, the smoother the curve or in other words, lower is the rate of decay because of the distance, as can be seen in Figure 2:

In some circumstances, the value of bandwidth b will add distinct characteristics of the basin under study, or the sample might interfere with the deposition and degradation of the pollutants in transit, such as the declivity, distribution of generation features, water flow, climate, molecular type, and weight, among other factors. Figure 3 presents a summary of the proposed model, its form of validation and application in the water resources planning and management.

Organic matter

One of the main indicators of anthropic activity in the water bodies used worldwide, is organic matter. This is because this compound can be found in waste that is caused by various polluting activities and when excessive, can either directly or indirectly, have serious effects on the equilibrium of the ecosystem. (Filella, 2009)

Nonetheless, organic matter is the basis of the whole food chain and participates in various natural processes since it is caused by natural activities such as the decomposition of vegetable waste and animal manure. Organic matter can be divided into two key factors - origin and composition - which are closely intertwined (Chen et al., 2002; Leenheer & Croué, 2003).

When natural organic matter originates outside the aquatic system, it is designated as allochthonous or pedogenic since it is largely made up of humic compounds which are refractory or in other words, have a slow rate of degradation owing to the complexity of their molecules and the fact that they have a high degree of aromaticity. However, the organic material produced inside the aquatic system through the decomposition of algae and excrement of aquatic biota or else through primary production, is called autochthonous or aquagenic and mainly has labile characteristics. This means they undergo rapid degradation owing to their lower molecular complexity which is predominantly aliphatic or allochthonous since thy have less molecular weight (Chen et al., 2002; Evans et al., 2005; Wilkinson et al., 1997; Zumstein & Buffle, 1989).

Owing to the differences in the rates of decomposition and in the density of the labile and refractory fractions of organic matter, it will be necessary to separate them since each one will follow different deposition/decomposition curves and thus have its own nucleus functions. Considering this, an essential stage required for the application of the model is to conduct the separation of these fractions. However, there is no robust technique for this as the principal method for determining the labile fraction of organic material is BOD, which is inconvenient in some ways, both regarding the time spent on their analysis and the permissible error rate that is involved.

Several authors have sought to characterize organic matter based on spectroscopic techniques, although principally through a qualitative method. Westerhoff and Anning (2000) analyzed samples of water bodies that had different hydrological characteristics and influenced domestic effluents, as well as samples of sewage effluent through UV-Visible absorption, for a wavelength of 254 nm, which was standardized by DOC, and designated specific ultraviolet absorbance in 254 nm (SUVA₂₅₄). On the basis of this comparison with a standard sample of refractory organic matter (fulvic acid from the River Suwannee), the authors noted that the greater values of SUVA₂₅₄ showed that the samples had a greater refractability, which was 4,4 L/(mg m) for the fulvic acid sample, whereas the lower values were linked to the samples of greater lability, with a value of 1.2 L/(mg m) for the sample of domestic effluent.

Weishaar et al. (2003) estimated the percentage of aromaticity in different samples of humic substances by means of C-13 Nuclear Magnetic Resonance and correlated with the SUVA₂₅₄ value of the samples, and obtained a function for calculating the percentage of aromaticity with R² = 0,97, as expressed in Equation 3:

In contrast, Her et al. (2008), assessed the SUVA₂₅₄ Index of refractory compounds, humic and fulvic acids, and the highly labile molecules, albumin and asparagine. They determined that the labile substances have values lower than SUVA (0.09 and 0.02 L/(mg m), respectively), whereas the humic substances had values of 7.35 and 4.21 L/(mg m), respectively, which confirmed that the more labile the compound, the lower the value of SUVA₂₅₄.

Based on these studies, it can be assumed that the aromaticity of the compound and hence its SUVA₂₅₄ Index can be used to estimate the degree of lability or refractability of a sample. Thus by applying Equation 3 to a value of SUVA₂₅₄ of 4.4 L/(mg m), relative to the fulvic acid, an aromaticity of 32.32% can be obtained which from this time will be the aromaticity value of the refractory substances. By noting that fulvic acid is a 100% refractory substance, one can apply a linear equation to obtain the percentage of refractory organic material (ROM) of the samples, as expressed in Equation 4:

Based on Equation 4, the percentage of refractory organic material can be obtained and the labile fraction, by noting the difference with the DOC. It should be stressed that this kind of separation can be used for the model set out here, even though this is not an exact way of separating the fractions of organic matter and further analyses needs to be conducted based on the sample compounds used by the known standard fractions of humic substances and proteins, since at this time there is no better separation process available.

Study Area

The site chosen for the application of the model is the main headwaters of the Barigui River basin, which is in the municipalities of Almirante Tamandaré and Colombo, in the State of Paraná, in the southern region of Brazil. The climate in the region is classified as Humid Subtropical (Cfb) in accordance with the Köppen classification system, without any well-defined dry season and a mild summer. The average annual rainfall is around 1,500 mm and the average temperature ranges from 16.1ºC to 18ºC (Froehner & Martins, 2008; Kozak, 2020).

The soil occupation in this part of the basin is predominantly rural with some agriculture and dense vegetation, although there are a few scattered urban communities. The main crops that are grown in the two municipalities are black beans and maize. Regarding the index of services provided through the sewage and drainage systems, according to the SNIS [National Sanitation Information System] of 2020 in Almirante Tamandaré, the coverage is approximately 58% and 74% in Colombo. However, according to the data provided by IBGE [Brazilian Institute of Geography and Statistics] and SNIS there is also a large treatment plant for individuals, which in 2010 was around 44% of the municipality of Almirante Tamandaré and 23% of the population of Colombo had septic tank sludge in their dwellings (IBGE, 2012; SNIS, 2021). Figure 4 and Table 1 show the location of the study area and the collection points.

Table 1

Location and drainage area of the collection points

Collection Point	x(m)	y (m)	Drainage Area (km²)
P1	671,633	7,200,115	43.08
P2	671,559	7,199,321	58.43

Source: Authors' elaboration

When obtaining input data from the model, water samples were collected along a stretch between two points of the main river of the Barigui River basin, in five sampling sessions (C1 to C5) on the following dates: 04/16 (autumn), 07/09 (winter) and 11/05 (spring) in 2018 and 02/18 (summer) and 05/20 (autumn) of 2019 to allow an assessment of the seasonal effects on organic matter. The climatic conditions, temperature and rainfall, on the dates of the collection are summarized at Figure 5. The red lines indicate the sampling dates.

The first collection (C1) was conducted in a period soon after a period of mild rainfall after a short dry season. The average temperature on the collection day was 15ºC, although on the preceding days it was higher and reached an average of 20ºC. C2 took place on a day when there was a sharp fall in temperature and the average on the preceding days was above 15ºC. It was also a period of drought when there had been little rainfall in the previous months.

In contrast, C3 took place after heavy rainfall followed by a short dry period. The average temperatures in the period shortly before the collection ranged from 15ºC to 20ºC. C4 was conducted soon after a rainfall of about 2 mm in a period of high temperatures and in the preceding days, the temperatures were above 20ºC. Finally, C5 took place in a period of constant rainfall but it was not very heavy, and the average temperatures ranged from 15ºC to 20ºC. All the collections took place during the morning. Table 2 shows the temperature data (average over a period of 3 days) and rainfall (total) for the 36h preceding the collections.

Table 2

Average Temperature and Total Precipitation in the previous 36 h before collection campaigns

Campaign	Average Temperature 36h (°C)	Total Precipitation 36h (mm)
C1	17.4	7.20
C2	17.3	0.00
C3	17.4	7.40
C4	20.5	60.40
C5	15.0	22.60

Source: National Institute of Meteorology (INMET), 2022

It is necessary to obtain the flow from the loading points at the time of the collection to calculate the load of organic matter at the collection points. However, there is only a fluviometric station at the P2 point. In view of this, the flow at this point was obtained from the depth of the river and its key curve. as shown in Equation 5 (Kozak, 2020).

where Q is the flow in m³/s and h the depth in cm. Following this, the flow at the P1 point was obtained by regularizing the flow conducted by means of the specific flow method. Table 3 displays the values at the level of the reference-point obtained from the data of the sampling sessions that were conducted.

Table 3

Reference-point levels at the time of collection

P2	C1	C2	C3	C4	C5
h (cm)	12.5	2	8	13	8

Source: Authors' elaboration with data from field

Based on the data obtained at the level of P2, the flows were calculated for this point by means of Equation 5, while the flows were calculated at Point P1 in a proportional way to their contributing areas, as was noted in Table 4.

Table 4

Flow rate at the time of collection

Collection Point	Drainage Area (km²)	C1 Flow (m³/s)	C2 Flow (m³/s)	C3 Flow (m³/s)	C4 Flow (m³/s)	C5 Flow (m³/s)
P1	43.08	0.41	0.19	0.31	0.42	0.31
P2	58.43	0.56	0.25	0.43	0.57	0.43

Source: Authors' elaboration with data from field and Equation 5 (Kozak, 2020)

Laboratory Parameters

After the collection, the samples were stored in glass amber flasks and then packed in thermal ice boxes until they reached the laboratory. The analysis of the DOC involved making use of approximately 50 mL of the samples filtered in cellulose acetate membrane with a porosity of 0.45 µm, and previously washed in ultra-pure water. Following this, it was transferred to an amber flask (calcined to 550°C) and acidified with 0.5% of the volume of PA sulphuric acid. Before the analyses, the acidified samples were purged with nitrogen gas to eliminate the dissolved inorganic carbon in the sample and thus avoid analytical errors. The DOC was measured in TOC V-CPH, Shimadzu brand equipment, through the high temperature combustion method (5310 B of Standard Methods) (APHA, 2005).

In the case of the analysis of UV-visible absorbance. about 4 mL of the filtered sample (but not acidified) was analyzed in a Spectrophotometer for Varian Absorbance (Cary 50 Conc model), in a quartz cell with 1 cm of an optical path. The reading of the absorbance spectrum was carried out between 200 and 600 nm with an interval of 1.0 nm and a scanning rate of 4800 nm/min. Ultrapure water was used like blank. The SUVA₂₅₄ Index was obtained to standardize the absorbance in 254 nm by the DOC and the correction by the optical path (l in m), having as the unit L/(mg m) (Westerhoff & Anning, 2000), as expressed in Equation 6.

The values of SUVA₂₅₄ were thus used for the separation of the fractions of organic matter and Equation 3 for calculating the aromaticity of the samples and following this, Equation 4 to obtain the concentration of refractory organic matter (ROM). The difference between the concentration of ROM and DOC, made it possible to obtain the concentration of labile organic matter (LOM). After this, the loads of ROM and LOM were obtained, based on the flows obtained at the time of the collection. Table 5 and Table 6 show the data referring to the C1 to C5 campaigns.

Table 5

Organic matter fractions at collection points

Campaign	Point	DOC (mg/L)	Absorbance 254 nm	SUVA₂₅₄ (L/(mg m))	% aromaticity	% ROM	% LOM
C1	P1	1,79	0,026410	1,47	13,24%	40,96%	59,04%
C1	P2	1,91	0,024093	1,26	11,87%	36,74%	63,26%
C2	P1	1,62	0,028361	1,75	15,04%	46,55%	53,45%
C2	P2	1,61	0,044374	2,76	21,60%	66,83%	33,17%
C3	P1	5,37	0,077288	1,44	13,01%	40,27%	59,73%
C3	P2	4,765	0,064227	1,35	12,42%	38,42%	61,58%
C4	P1	4,258	0,110656	2,60	20,57%	63,66%	36,34%
C4	P2	4,106	0,126874	3,09	23,78%	73,57%	26,43%
C5	P1	3,247	0,066275	2,04	16,94%	52,41%	47,59%
C5	P2	3,601	0,041710	1,16	11,18%	34,60%	65,40%

Table 6

Refractory and labile organic matter loads

Campaign	Point	DOC (mg/L)	% ROM	% LOM	ROM (mg/L)	LOM (mg/L)	Flow (m³/s)	Load ROM (kg/day)	Load LOM (kg/day)
C1	P1	1.79	40.96%	59.04%	0.73	1.06	0.4113	26	38
C1	P2	1.91	36.74%	63.26%	0.70	1.21	0.5578	34	58
C2	P1	1.62	46.55%	53.45%	0.75	0.87	0.1874	12	14
C2	P2	1.61	66.83%	33.17%	1.08	0.53	0.2541	24	12
C3	P1	5.37	40.27%	59.73%	2.16	3.21	0.3136	59	87
C3	P2	4.765	38.42%	61.58%	1.83	2.93	0.4253	67	108
C4	P1	4.258	63.66%	36.34%	2.71	1.55	0.4223	99	56
C4	P2	4.106	73.57%	26.43%	3.02	1.09	0.5727	149	54
C5	P1	3.247	52.41%	47.59%	1.70	1.55	0.3136	46	42
C5	P2	3.601	34.60%	65.40%	1.25	2.36	0.4253	46	87

Satellite images

The land cover can be obtained through the classification of satellite images by means of GIS tools. Thus, when correlating the data on the organic matter obtained in the collections, satellite images were selected on the dates closest to the collections conducted. The satellite chosen was Sentinel 2 which has free images and a high spectral resolution since the bands in the visible and ultraviolet range were close to 10 m (bands 2 – blue, 3 – green, 4 – red, 8 – close to infrared), and 20 m for the red edge bands (5, 6, 7 and 8A bands) and short waves (11 and 12 bands) (USGS, 2021).

The study watershed is included in the T22JFT scene of the Sentinel-2 series. The data from the images were chosen (provided the images were available with few clouds) from the site of the United States Geological Service (USGS), in periods close to the time when the collections were conducted. After the analysis of the current images, the data obtained were 05/15, 07/09 and 09/07 in 2018 and 02/09 and 06/09 in 2019. The shape of the Barigüi River basin was acquired at the IAT (Water and Land Institute) site (IAT, 2020). However, the sub-basins contributed by the points in the analysis, were obtained by means of the contour lines available in the AMBDATA program (INPE, 2021) for the region in question.

The projection system used for the images was the same shape as the basin that was made available – SAD 69 UTM 22S. The bands were stacked to make it possible to proceed with the classification of the soil cover and following this, the cutting of the sub-basin was conducted in the shape of the basin formed by the shape of the P1 and P2 points. These stages were prepared in QGIS free software. Figure 6 presents the images used in the land cover classification.

The classes of coverage were determined by analyzing a file with stacked bands in MultiSpec software by employing different combinations of bands available. Each set of combinations highlights certain features of the classes. Together with this, images from Google Earth Pro and Street View were looked at so that the visualization of a better coverage resolution could be obtained in their analysis and the class of each sample that belonged to it could be determined. The choice of samples took account of the fact that a uniform analysis would be required for all the types of coverage found in study area. Thus, the classification of the images was supervised, and the following classes of coverage were identified: Forest (fo), Grass (gr), Agriculture (ag), Bared Land (bl), Urban Area (ua) and Water (wa). Figure 7 shows the classified images.

After this supervised classification, images were created with a much higher number of features including tiny areas that were not needed for the designed model (for example of a pixel, or 100 m²). Hence, the features with an area less than a 0,026 km² (which is equivalent to 0.05% of the study basin) were combined with the adjoining features to make it easier to conduct the subsequent data processing of the existing coverage. At the same time, any corrections regarded as necessary, were automatically made to the classified features. This process was aided by comparing the features obtained with the images of Google Earth for the next dates available. The final classification of the soil cover for the area under study can be found in Figure 8.

The LUPC 1.0 program

The model was put into effect by creating a Python language code (shown in the Supplementary Material). The model was then employed for the data described earlier to ensure its applicability, as well as allowing the PCI parameters to be calculated for the several types of land cover analyzed and the bandwidth b regarding each collection.

When determining the PCI and b values, sets of PCI values were applied to each type of landcover with a pre-established range of values (each PCI ranging from 1 to 10, for example). It also involved applying Equation 1 with the resulting value equal to the load at the monitoring point of the river and the Gaussian (Kernel) function (Equation 2) to determine the weight of each feature. The value of b was found by applying the bisection method to determine the roots of an equation. The set of values that obtains the lowest error show the PCI and b values that are most suitable for this session.

The calculation was made for a load of 10² g/day because the program only uses whole numbers of PCI with a view to improving the precision of the values obtained. However, the load is displayed in terms of kg/day and PCI in kg/(day km²), which are the values calculated for the PCI to one decimal place.

LUCP Model Validation

At the end of the calculation for the first four sessions, a set of PCI values and b was obtained for each collection, which represent the four seasons of the year. The validation of the model with data from the C5 session made use of both the PCI and b values obtained for the same season (autumn), as well as the average of the values obtained in the four seasons for verification. At the same time, the weights for each feature were calculated based on the average bandwidth b that was found, and the loads produced by ROM and LOM were calculated. They were then added to the product of each load through its equivalent weight to obtain the load values of ROM and LOM for the P1 and P2 points. These values were compared with the real-world values of the load available, and the errors were calculated to evaluate the applicability of the model on screen.

Table 7 and Table 8 and Figure 9 and Figure 10 show the results of the model for PCI and b to refractory and labile organic matter, respectively.

Table 7

PCI and b for refractory organic matter

Camp.	PCI ROM gr	PCI ROM fo	PCI ROM ag	PCI ROM ua	PCI ROM bl	PCI ROM wa	b (m)	ROM Load P1 (kg/day)	Error (log)
C1 (aut)	12.9	4.2	7	1.7	14.2	17.9	1.6621	26.0	-8
C2 (win)	1.8	12.1	2	13	5.1	7.1	1.1081	12.0	-7
C3 (spr)	52.2	7	43.9	6.3	47.4	36.9	1.3519	59.0	-8
C4 (sum)	9	8	12.6	8	3.7	4.9	2.5327	99.0	-10
Average	19.0	7.8	16.4	7.3	17.6	16.7	1.6637
SD	22.6	3.3	18.9	4.7	20.4	14.6	0.6221

Table 8

PCI and b for labile organic matter

Camp.	PCI LOM gr	PCI LOM fo	PCI LOM ag	PCI LOM ua	PCI LOM bl	PCI LOM wa	b (m)	LOM Load P1 (kg/day)	Error (log)
C1 (aut)	11.8	5.8	10.7	9.2	9.2	7.8	1.7517	38.00	-7
C2 (win)	16.1	1	15.6	1.4	21.9	21.4	1.0747	14.00	-9
C3 (spr)	16.9	1	17.8	4.9	1.9	19.1	4.0365	87.00	-10
C4 (sum)	22	27	212	14	10	162	1.0682	56.00	-10
Average	16.7	8.7	64.0	7.4	10.8	52.6	1.9827
SD	4.2	12.4	98.7	5.4	8.3	73.2	1.4062

As can be determined, the highest values of PCI ROM were obtained for the coverage of grass (gr), agriculture (ag), bared land (bl) and water (wa) for the spring (C3). This can be explained by the heavy rainfall that occurred a few days before the collection which led to a lot of material being carried to the inside of the river. As this heavy downpour was followed by a period of dry days, the labile organic matter could have been degraded during this period and left the main part of the refractory fraction remaining in the river.

In contrast, the LOM showed high values for the collection conducted in summer (C4) for the forest, agricultural and water coverage. This collection was undertaken in a period of high temperatures which favored the primary production in the lakes of the basin (high PCI wa) as well as the soil, or in other words, in the general areas of vegetation (PCI ra, ar and ag). In addition, soon after the collection, there was a rainfall of about 2 mm which led to this organic material being dragged to the river. The average values both for the LOM, and ROM, were seriously affected by these two events, which is noticeable in Figures 9 and 10 where the other values of PCI show less variation.

About the PCI values for urban areas, a lower variation was found for both the ROM and LOM. This could be expected because there is not much influence of seasonal climatic factors on the organic matter load which is mainly produced by domestic waste. As well as this, the urban influence was not very high because the basin offers a good cover for the treatment of sewage, whether it is collective or individual. Moreover, when dealing with a region that is essentially rural, the dwellings are more scattered and mainly consist of single families and thus have less impact on large urban agglomerations.

When noting the PCI values that were found, it was also determined that the areas of vegetation make a predominantly refractory contribution (the tree vegetation had an average of LOM higher than ROM, although the values of the latter showed greater variability), as well as a bared land, where there was a more rapid loss of macronutrients which tended to lead to sediments that contained a higher proportion of ROM. On the other hand, the lakes make a greater labile contribution or in other words, have aquagenic organic matter which shows greater lability. Thus, the values for urban areas are similar, with a slightly greater proportion of LOM. However, when one thinks of other measurement parameters of organic matter, which is necessary in the case of sewage waste disposal, the BOD generally is 50% of COD value, or put more simply, of total organic matter where about half refers to LOM, which corroborates the values that were found.

Considering this, the PCI values found by the model for the available data are in accordance with the climatic, temperature and rainfall features in the period of the collections and with the already known characteristics of organic matter based on their origin. Thus, it is now necessary to assess the values of the bandwidth b, or simply the way organic matter decays.

In some circumstances, b includes several features that interfere with the kind of degradation and deposition of the pollutants. In this case, the organic matter, which is different for the ROM and LOM, has a greater molecular weight, since the refractory fraction in most situations, only undergoes deposition, whereas the labile fraction is also susceptible to photo- and biodegradation. As can be seen, the b had a greater variability for the LOM, which was expected because apart from the deposition, there is a degradation factor added to this fraction, which is also influenced by a few factors such as insolation, temperature, and the existing bacterial community.

Correlations were applied to determine which parameters influence the b values for both LOM and ROM. The purpose of this was to understand the main influences on the b values among the parameters available for the data collections i.e., the average temperature, and total rainfall for the last three days. Figure 11 and Figure 12 show the correlations found between these parameters and their coefficient of determination (R²).

As seen, the ROM had a significant positive correlation about temperature and rainfall (R² = 0.8913 and 0.923, respectively), whereas the LOM had a low negative correlation regarding these parameters (R² = 0.1652 and 0.1249, respectively). The behavioral pattern of b for the ROM lives up to expectations, since the drag caused by this rainfall compound must be one of the key factors in its deposition along the path, as far as the monitoring points.

This is because the greater the dragging speed, the lower the deposition of material and hence results in a greater bandwidth b. As regards the influence of temperature, this can be better explained because of the close correlation with the rainfall parameter (R² = 0.9939), since there was no clear direct relationship between the temperature on the collection day and the deposition itself.

In the case of the LOM, it is expected that more factors will be needed to explain the behavioral pattern of b, owing to the greater added complexity and the fact that apart from the deposition, there is no degradation along the trajectory. Moreover, there is no close correlation between the LOM b and any factor such as is the case with ROM. Regarding the parameters analyzed here, both temperature and rainfall, there is a small negative correlation with LOM b (R² = 0.8913 and 0.923, respectively). This is because the highest temperatures accelerate the process of biodegradation of the ROM, increase its decay and hence reduce the value of b. Regarding rainfall, this exerts a force on the dragging of the ROM along the basin to the inside of the water bodies and causes less interference to the deposition than to the ROM because it has less molecular weight than LOM.

Finally, a fifth collection was conducted on 05/20/2020, during the autumn, to confirm the validity of the model for determining the pollutant load in the river based on the PCI data that was found for the ROM and LOM. Thus, the validation took place in two ways and began by applying the values obtained in the first collection in autumn for a model with a seasonal application and an assessment of the predictive capacity of the load for the monitoring points (P1 and P2). However, the measured data of the values found were also applied and envisaged an application on a global scale to understand the potential danger of the pollution in the basin, regardless of the season of the year. Table 9 shows the data obtained for the values of PCI and b obtained for autumn and the average values, as well as the errors that were made in each application.

Table 9

Organic matter estimation by LUPC model

PCI Season	Organic Matter	Point	Estimated Load (kg/day)	Measured Load (kg/day)	Error	Average Error
Autumn	Refractory	P1	27.73	46	39.71%	21.52%
	Refractory	P2	37.16	46	19.22%
	Labile	P1	41.90	42	0.24%
	Labile	P2	63.59	87	26.90%
Average	Refractory	P1	49.36	46	7.31%	60.84%
	Refractory	P2	70.60	46	53.48%
	Labile	P1	95.59	42	127.59%
	Labile	P2	134.82	87	54.97%

As can be noted from the data shown, the seasonal application provided a suitable estimate of the organic matter load for the hydrographic basin under study, with an average error rate that was lower than 22%, which is approximately the level of errors that are acceptable to the BOD, for example. Thus, the application of the model can be used as a complementary device or even a substitute for the constant monitoring sessions that incur excessive costs. About the application of the average values, the error was greater and almost 61%. However, the purpose of this application is to assess the potential danger of pollution, with a greater number of errors being acceptable in this case.

Regardless of how it is used, whether in a seasonal or global way, the model has proved to be useful for estimating the organic matter - both labile and refractory - based on the soil cover of the basin under study. Thus, it can be used as an alternative to the well-known conventional monitoring of water quality. As well as this, the model can be used as a tool to assist decision-making because it can be applied in strategic situations that involve integrated planning between urban and rural soil for the management of water resources. This is because after making small alterations in the maps that provide input data for soil cover, situations can be foreseen where new gated communities are established or areas of vegetation are turned into pastureland, for example.

Another possible application of the model is in the idea of improving the locality by taking measures aimed at depollution. such as sewage treatment plants and assessing the difference in the pollutant loads. This should be conducted in accordance with the removal of widespread drainage systems in the region and incorporating a point of waste treatment in the model. As well as this, the model can be easily altered to include other parameters such as nutrients, alternative determination parameters of organic matter, metals, and emerging compounds.

Because of what is examined here, the LUPC model has proved to be applicable for the management of water resources, either by replacing or adding to the monitoring of water quality or as a system that can aid decision-making.

The study undertaken here sets out a scheme for the LUPC model and its parameters and envisages assessing the dangers of pollution in a river with a soil cover at its hydrographic basin. It also involved a means of separating labile and refractory fractions of organic matter by making use of the SUVA₂₅₄ Index and the aromaticity of the sample. As seen, the model proved to be applicable, and the PCI and b values were defined for different seasons of the year in the hydrographic basin in the study of the organic matter parameters.

In future studies, the LUPC model can be applied to other basins to determine the PCI values found and the factors that affect the value of the bandwidth b, particularly about the labile fraction of the organic matter, as well as applying the model to more monitoring points in other situations such as semi-arid regions.

Regarding the separation of organic matter, it also shows how the equation used for blended patterns of humic and labile substances of a known concentration, can be employed to assess its efficiency. A last point is the application of the model for other parameters of water quality and to define how they can be combined.

Ahmad, S. R., & Reynolds, D. M. (1999). Monitoring of water quality using fluorescence technique: prospect of on-line process control. Water Research, 33(9), 2069–2074. https://doi.org/10.1016/S0043-1354(98)00435-7
APHA. (2005). Standard Methods for the Examination of Water and Wastewater Standard Methods for the Examination of Water and Wastewater. 1, 1–14. https://doi.org/10.2105/AJPH.51.6.940-a
Asming, M. A. A., Ibrahim, A. M., & Abir, I. M. (2022). Processing and classification of landsat and sentinel images for oil palm plantation detection. Remote Sensing Applications: Society and Environment, 26(June 2021), 100747. https://doi.org/10.1016/j.rsase.2022.100747
Bostanmaneshrad, F., Partani, S., Noori, R., Nachtnebel, H. P., Berndtsson, R., & Adamowski, J. F. (2018). Relationship between water quality and macro-scale parameters (land use, erosion, geology, and population density) in the Siminehrood River Basin. Science of the Total Environment, 639, 1588–1600. https://doi.org/10.1016/j.scitotenv.2018.05.244
Carstea, E. M. (2012). Fluorescence Spectroscopy as a Potential Tool for In-Situ Monitoring of Dissolved Organic Matter in Surface Water Systems. In N. Balkis (Ed.), Water Pollution (pp. 47–68). InTech. http://cdn.intechopen.com/pdfs/28946/InTech-Fluorescence_spectroscopy_as_a_potential_tool_for_in_situ_monitoring_of_dissolved_organic_matter_in_surface_water_systems.pdf
Carstea, E. M., Ghervase, L., Pavelescu, G., & Iojă, C. I. (2012). Correlation of Dissolved Organic Matter Fluorescence and Several Metals Concentration in a Freshwater System. Procedia Environmental Sciences, 14, 41–48. https://doi.org/10.1016/j.proenv.2012.03.005
Chen, J., Gu, B., LeBoeuf, E. . J., Pan, H., DAI, S., BAOHUA, G., LeBoeuf, E. . J., HONGJUN, P., & DAI, S. (2002). Spectroscopic characterization of the structural and functional properties of natural organic matter fractions , . Chemosphere, 48(2002), 59–68.
de Mesnard, L. (2013). Pollution models and inverse distance weighting: Some critical remarks. Computers and Geosciences, 52, 459–469. https://doi.org/10.1016/j.cageo.2012.11.002
Edzwald, J. K., Becker, C. W., L., W. K., Journal, S., Water, A., Association, W., April, R. T., Edzwald, J. K., Becker, W. C., & Wattier, K. L. (1985). American Water Works Association, 77(4), 122–132.
Evans, C. D., Monteith, D. T., & Cooper, D. M. (2005). Long-term increases in surface water dissolved organic carbon: Observations, possible causes and environmental impacts. Environmental Pollution, 137(1), 55–71. https://doi.org/10.1016/j.envpol.2004.12.031
Ferreira, D. B., Mühlenhoff, A. P., & Fernandes, C. V. S. (2018). Modelos de poluição difusa: desafios, estratégias e impacto para a gestão de recursos hídricos. Revista de Gestão de Água da América Latina, 15(2018).
Filella, M. (2009). Freshwaters: Which NOM matters? Environmental Chemistry Letters, 7(1), 21–35. https://doi.org/10.1007/s10311-008-0158-x
Froehner, S., & Martins, R. F. (2008). Avaliação da composição química de sedimentos do Rio Barigüi na Região Metropolitana de Curitiba. Química Nova, 31(8), 2020–2026.
Giri, S., & Qiu, Z. (2016). Understanding the relationship of land uses and water quality in Twenty First Century: A review. Journal of Environmental Management, 173, 41–48. https://doi.org/10.1016/j.jenvman.2016.02.029
Her, N., Amy, G., Sohn, J., & Gunten, U. (2008). UV absorbance ratio index with size exclusion chromatography (URI-SEC) as an NOM property indicator. Journal of Water Supply: Research and Technology - AQUA, 57(1), 35–44. https://doi.org/10.2166/aqua.2008.029
Huang, J., Huang, Y., Pontius, R. G., & Zhang, Z. (2015). Geographically weighted regression to measure spatial variations in correlations between water pollution versus land use in a coastal watershed. Ocean and Coastal Management, 103, 14–24. https://doi.org/10.1016/j.ocecoaman.2014.10.007
Hudson, N., Baker, A., Ward, D., Reynolds, D. M., Brunsdon, C., Carliell-Marquet, C., & Browning, S. (2008). Can fluorescence spectrometry be used as a surrogate for the Biochemical Oxygen Demand (BOD) test in water quality assessment? An example from South West England. Science of The Total Environment, 391(1), 149–158. https://doi.org/10.1016/j.scitotenv.2007.10.054
Hur, J., & Kong, D. ‐S. (2008). Use of Synchronous Fluorescence Spectra To Estimate Biochemical Oxygen Demand (Bod) of Urban Rivers Affected By Treated Sewage. Environmental Technology, 29(4), 435–444. https://doi.org/10.1080/09593330801984472
IAT – INSTITUTO ÁGUA E TERRA (2020) Mapas e Dados Espaciais. Available from: https://www.iat.pr.gov.br/Pagina/Mapas-e-Dados-Espaciais. Acessed on: Feb 15th, 2020.
IBGE – INSTITUTO BRASILEIRO DE GEOGRAFIA E ESTATÍSTICA (2012). Censo Brasileiro de 2010. Rio de Janeiro: IBGE
INMET – INSTITUTO NACIONAL DE METEOROLOGIA (2022). Gráficos Anuais de Estações Automáticas. Curitiba. Available from: https://tempo.inmet.gov.br/GraficosAnuais/A807. Acessed on: Apr 30th, 2022.
INPE – INSTITUTO NACIONAL DE PESQUISAS ESPACIAIS (2020) Variáveis Ambientais para Modelagem de distribuição de espécies. AMBDATA. Available from: http://www.dpi.inpe.br/ Ambdata/ unidades_administrativas.php. Acessed on: Feb 09th, 2020.
Jouanneau, S., Recoules, L., Durand, M. J., Boukabache, a., Picot, V., Primault, Y., Lakel, a., Sengelin, M., Barillon, B., & Thouand, G. (2014). Methods for assessing biochemical oxygen demand (BOD): A review. Water Research, 49(1), 62–82. https://doi.org/10.1016/j.watres.2013.10.066
Koh, E. H., Lee, E., & Lee, K. K. (2020). Application of geographically weighted regression models to predict spatial characteristics of nitrate contamination: Implications for an effective groundwater management strategy. Journal of Environmental Management, 268, 110646. https://doi.org/10.1016/j.jenvman.2020.110646
Kozak, C. (2020). Non point source pollution assessment based on event driven influence. (Doctoral Dissertation, Federal University of Parana, Brazil). Federal University of Parana Research Repository. https://www.acervodigital.ufpr.br/handle/1884/70367
Kumar, A., Kanga, S., Taloor, A. K., Singh, S. K., & Đurin, B. (2021). Surface runoff estimation of Sind river basin using integrated SCS-CN and GIS techniques. HydroResearch, 4, 61–74. https://doi.org/10.1016/j.hydres.2021.08.001
Kumar, R.; Kumar, A. Water Analysis: Biochemical Oxygen Demand. In: Worsfold, P.; Townshend, A.; Poole C., eds. Encyclopedia of Analytical Science. 2 ed. Amsterdam: Elsevier Academic Press. p.315 – 325. 2005.
Leenheer, J. A., & Croué, J.-P. (2003). Peer reviewed: characterizing aquatic dissolved organic matter. Environmental Science & Technology, 37(1), 18A-26A. https://doi.org/10.1021/es032333c
Li, S., Zhang, J., Guo, E., Zhang, F., Ma, Q., & Mu, G. (2017). Dynamics and ecological risk assessment of chromophoric dissolved organic matter in the Yinma River Watershed: Rivers, reservoirs, and urban waters. Environmental Research, 158(February), 245–254. https://doi.org/10.1016/j.envres.2017.06.020
Mir, A., Piri, J., & Kisi, O. (2017). Spatial monitoring and zoning water quality of Sistan River in the wet and dry years using GIS and geostatistics. Computers and Electronics in Agriculture, 135, 38–50. https://doi.org/10.1016/j.compag.2017.01.022
Mishra, S. K., & Singh, V. P. (1999). Another Look at SCS-CN Method. Journal of Hydrologic Engineering, 4(3), 257–264.
Peuravuori, J., Koivikko, R., & Pihlaja, K. (2002). Characterization , differentiation and classification of aquatic humic matter separated with different sorbents: synchronous scanning fluorescence spectroscopy. Water Research, 36, 4552–4562.
Rimba, A. B., Mohan, G., Chapagain, S. K., Arumansawang, A., Payus, C., Fukushi, K., Husnayaen, Osawa, T., & Avtar, R. (2021). Impact of population growth and land use and land cover (LULC) changes on water quality in tourism-dependent economies using a geographically weighted regression approach. Environmental Science and Pollution Research. https://doi.org/10.1007/s11356-020-12285-8
Rostan, J. C., & Cellot, B. (1995). On the use of U V spectrophotometry to assess dissolved organic carbon origin variations in the Upper Rhone River. Aquatic Sciences, 57(1), 70–80.
SNIS – Sistema Nacional de Informação sobre Saneamento (2021). Diagnóstico dos serviços de Água e Esgoto - 2020 Brasília: SNIS.
Trujillo-Jiménez, M. A., Liberoff, A. L., Pessacg, N., Pacheco, C., Díaz, L., & Flaherty, S. (2022). SatRed: New classification land use/land cover model based on multi-spectral satellite images and neural networks applied to a semiarid valley of Patagonia. Remote Sensing Applications: Society and Environment, 26(January). https://doi.org/10.1016/j.rsase.2022.100703
Tu, J., & Xia, Z. G. (2008). Examining spatially varying relationships between land use and water quality using geographically weighted regression I: Model design and evaluation. Science of the Total Environment, 407(1), 358–378. https://doi.org/10.1016/j.scitotenv.2008.09.031
USGS – UNITED STATES GEOLOGICAL SURVEY (2020). Earth Explorer. Available from: https://earthexplorer.usgs.gov/. Acessed on: Jan 20th, 2020.
Visco, G., Campanella, L., & Nobili, V. (2005). Organic carbons and TOC in waters: An overview of the international norm for its measurements. Microchemical Journal, 79(1–2), 185–191. https://doi.org/10.1016/j.microc.2004.10.018
Walega, A., Amatya, D. M., Caldwell, P., Marion, D., & Panda, S. (2020). Assessment of storm direct runoff and peak flow rates using improved SCS-CN models for selected forested watersheds in the Southeastern United States. Journal of Hydrology: Regional Studies, 27(August 2019), 100645. https://doi.org/10.1016/j.ejrh.2019.100645
Wang, X., & Zhang, F. (2018). Effects of land use/cover on surface water pollution based on remote sensing and 3D-EEM fluorescence data in the Jinghe Oasis. Scientific Reports, 8(1), 1–13. https://doi.org/10.1038/s41598-018-31265-0
Weishaar, J. L., Fram, M. S., Fujii, R., & Mopper, K. (2003). Evaluation of Specific Ultraviolet Absorbance as an Indicator of the Chemical Composition and Reactivity of Dissolved Organic Carbon. 4702–4708. https://doi.org/10.1021/es030360x
Westerhoff, P., & Anning, D. (2000). Concentrations and characteristics of organic carbon in surface water in Arizona: Influence of urbanization. Journal of Hydrology, 236(3–4), 202–222. https://doi.org/10.1016/S0022-1694(00)00292-4
Wilkinson, K. J., Joz-Roland, A., & Buffle, J. (1997). Different roles of pedogenic fulvic acids and aquagenic biopolymers on colloid aggregation and stability in freshwaters. Limnology and Oceanography, 42, 1714–1724. https://doi.org/10.4319/lo.1997.42.8.1714
Wilson, C. O. (2015). Land use/land cover water quality nexus: quantifying anthropogenic influences on surface water quality. Environmental Monitoring and Assessment, 187(7). https://doi.org/10.1007/s10661-015-4666-4
Zhang, Q., Shen, J., & Sun, F. (2021). Spatiotemporal differentiation of coupling coordination degree between economic development and water environment and its influencing factors using GWR in China’s province. Ecological Modelling, 462(September), 109794. https://doi.org/10.1016/j.ecolmodel.2021.109794
Zhao, W., Zhu, X., Sun, X., Shu, Y., & Li, Y. (2015). Water quality changes in response to urban expansion: spatially varying relations and determinants. Environmental Science and Pollution Research, 22(21), 16997–17011. https://doi.org/10.1007/s11356-015-4795-x
Zumstein, J., & Buffle, J. (1989). Circulation of pedogenic and aquagenic organic matter in an eutrophic lake. Water Research, 23(2), 229–239. https://doi.org/10.1016/0043-1354(89)90047-X

Ahmad, S. R., & Reynolds, D. M. (1999). Monitoring of water quality using fluorescence technique: prospect of on-line process control. Water Research, 33(9), 2069–2074. https://doi.org/10.1016/S0043-1354(98)00435-7
APHA. (2005). Standard Methods for the Examination of Water and Wastewater Standard Methods for the Examination of Water and Wastewater. 1, 1–14. https://doi.org/10.2105/AJPH.51.6.940-a
Asming, M. A. A., Ibrahim, A. M., & Abir, I. M. (2022). Processing and classification of landsat and sentinel images for oil palm plantation detection. Remote Sensing Applications: Society and Environment, 26(June 2021), 100747. https://doi.org/10.1016/j.rsase.2022.100747
Bostanmaneshrad, F., Partani, S., Noori, R., Nachtnebel, H. P., Berndtsson, R., & Adamowski, J. F. (2018). Relationship between water quality and macro-scale parameters (land use, erosion, geology, and population density) in the Siminehrood River Basin. Science of the Total Environment, 639, 1588–1600. https://doi.org/10.1016/j.scitotenv.2018.05.244
Carstea, E. M. (2012). Fluorescence Spectroscopy as a Potential Tool for In-Situ Monitoring of Dissolved Organic Matter in Surface Water Systems. In N. Balkis (Ed.), Water Pollution (pp. 47–68). InTech. http://cdn.intechopen.com/pdfs/28946/InTech-Fluorescence_spectroscopy_as_a_potential_tool_for_in_situ_monitoring_of_dissolved_organic_matter_in_surface_water_systems.pdf
Carstea, E. M., Ghervase, L., Pavelescu, G., & Iojă, C. I. (2012). Correlation of Dissolved Organic Matter Fluorescence and Several Metals Concentration in a Freshwater System. Procedia Environmental Sciences, 14, 41–48. https://doi.org/10.1016/j.proenv.2012.03.005
Chen, J., Gu, B., LeBoeuf, E.. J., Pan, H., DAI, S., BAOHUA, G., LeBoeuf, E.. J., HONGJUN, P., & DAI, S. (2002). Spectroscopic characterization of the structural and functional properties of natural organic matter fractions,. Chemosphere, 48(2002), 59–68.
de Mesnard, L. (2013). Pollution models and inverse distance weighting: Some critical remarks. Computers and Geosciences, 52, 459–469. https://doi.org/10.1016/j.cageo.2012.11.002
Edzwald, J. K., Becker, C. W., L., W. K., Journal, S., Water, A., Association, W., April, R. T., Edzwald, J. K., Becker, W. C., & Wattier, K. L. (1985). American Water Works Association, 77(4), 122–132.
Evans, C. D., Monteith, D. T., & Cooper, D. M. (2005). Long-term increases in surface water dissolved organic carbon: Observations, possible causes and environmental impacts. Environmental Pollution, 137(1), 55–71. https://doi.org/10.1016/j.envpol.2004.12.031
Ferreira, D. B., Mühlenhoff, A. P., & Fernandes, C. V. S. (2018). Modelos de poluição difusa: desafios, estratégias e impacto para a gestão de recursos hídricos. Revista de Gestão de Água da América Latina, 15(2018).
Filella, M. (2009). Freshwaters: Which NOM matters? Environmental Chemistry Letters, 7(1), 21–35. https://doi.org/10.1007/s10311-008-0158-x
Froehner, S., & Martins, R. F. (2008). Avaliação da composição química de sedimentos do Rio Barigüi na Região Metropolitana de Curitiba. Química Nova, 31(8), 2020–2026.
Giri, S., & Qiu, Z. (2016). Understanding the relationship of land uses and water quality in Twenty First Century: A review. Journal of Environmental Management, 173, 41–48. https://doi.org/10.1016/j.jenvman.2016.02.029
Her, N., Amy, G., Sohn, J., & Gunten, U. (2008). UV absorbance ratio index with size exclusion chromatography (URI-SEC) as an NOM property indicator. Journal of Water Supply: Research and Technology - AQUA, 57(1), 35–44. https://doi.org/10.2166/aqua.2008.029
Huang, J., Huang, Y., Pontius, R. G., & Zhang, Z. (2015). Geographically weighted regression to measure spatial variations in correlations between water pollution versus land use in a coastal watershed. Ocean and Coastal Management, 103, 14–24. https://doi.org/10.1016/j.ocecoaman.2014.10.007
Hudson, N., Baker, A., Ward, D., Reynolds, D. M., Brunsdon, C., Carliell-Marquet, C., & Browning, S. (2008). Can fluorescence spectrometry be used as a surrogate for the Biochemical Oxygen Demand (BOD) test in water quality assessment? An example from South West England. Science of The Total Environment, 391(1), 149–158. https://doi.org/10.1016/j.scitotenv.2007.10.054
Hur, J., & Kong, D. -S. (2008). Use of Synchronous Fluorescence Spectra To Estimate Biochemical Oxygen Demand (Bod) of Urban Rivers Affected By Treated Sewage. Environmental Technology, 29(4), 435–444. https://doi.org/10.1080/09593330801984472
IAT – INSTITUTO ÁGUA E TERRA (2020) Mapas e Dados Espaciais. Available from: https://www.iat.pr.gov.br/Pagina/Mapas-e-Dados-Espaciais. Acessed on: Feb 15th, 2020.
IBGE – INSTITUTO BRASILEIRO DE GEOGRAFIA E ESTATÍSTICA (2012). Censo Brasileiro de 2010. Rio de Janeiro: IBGE
INMET – INSTITUTO NACIONAL DE METEOROLOGIA (2022). Gráficos Anuais de Estações Automáticas. Curitiba. Available from: https://tempo.inmet.gov.br/GraficosAnuais/A807. Acessed on: Apr 30th, 2022.
INPE – INSTITUTO NACIONAL DE PESQUISAS ESPACIAIS (2020) Variáveis Ambientais para Modelagem de distribuição de espécies. AMBDATA. Available from: http://www.dpi.inpe.br/Ambdata/ unidades_administrativas.php. Acessed on: Feb 09th, 2020.
Jouanneau, S., Recoules, L., Durand, M. J., Boukabache, a., Picot, V., Primault, Y., Lakel, a., Sengelin, M., Barillon, B., & Thouand, G. (2014). Methods for assessing biochemical oxygen demand (BOD): A review. Water Research, 49(1), 62–82. https://doi.org/10.1016/j.watres.2013.10.066
Koh, E. H., Lee, E., & Lee, K. K. (2020). Application of geographically weighted regression models to predict spatial characteristics of nitrate contamination: Implications for an effective groundwater management strategy. Journal of Environmental Management, 268, 110646. https://doi.org/10.1016/j.jenvman.2020.110646
Kozak, C. (2020). Non point source pollution assessment based on event driven influence. (Doctoral Dissertation, Federal University of Parana, Brazil). Federal University of Parana Research Repository. https://www.acervodigital.ufpr.br/handle/1884/70367
Kumar, A., Kanga, S., Taloor, A. K., Singh, S. K., & Đurin, B. (2021). Surface runoff estimation of Sind river basin using integrated SCS-CN and GIS techniques. HydroResearch, 4, 61–74. https://doi.org/10.1016/j.hydres.2021.08.001
Kumar, R.; Kumar, A. Water Analysis: Biochemical Oxygen Demand. In: Worsfold, P.; Townshend, A.; Poole C., eds. Encyclopedia of Analytical Science. 2 ed. Amsterdam: Elsevier Academic Press. p.315–325. 2005.
Leenheer, J. A., & Croué, J.-P. (2003). Peer reviewed: characterizing aquatic dissolved organic matter. Environmental Science & Technology, 37(1), 18A-26A. https://doi.org/10.1021/es032333c
Li, S., Zhang, J., Guo, E., Zhang, F., Ma, Q., & Mu, G. (2017). Dynamics and ecological risk assessment of chromophoric dissolved organic matter in the Yinma River Watershed: Rivers, reservoirs, and urban waters. Environmental Research, 158(February), 245–254. https://doi.org/10.1016/j.envres.2017.06.020
Mir, A., Piri, J., & Kisi, O. (2017). Spatial monitoring and zoning water quality of Sistan River in the wet and dry years using GIS and geostatistics. Computers and Electronics in Agriculture, 135, 38–50. https://doi.org/10.1016/j.compag.2017.01.022
Mishra, S. K., & Singh, V. P. (1999). Another Look at SCS-CN Method. Journal of Hydrologic Engineering, 4(3), 257–264.
Peuravuori, J., Koivikko, R., & Pihlaja, K. (2002). Characterization, differentiation and classification of aquatic humic matter separated with different sorbents: synchronous scanning fluorescence spectroscopy. Water Research, 36, 4552–4562.
Rimba, A. B., Mohan, G., Chapagain, S. K., Arumansawang, A., Payus, C., Fukushi, K., Husnayaen, Osawa, T., & Avtar, R. (2021). Impact of population growth and land use and land cover (LULC) changes on water quality in tourism-dependent economies using a geographically weighted regression approach. Environmental Science and Pollution Research. https://doi.org/10.1007/s11356-020-12285-8
Rostan, J. C., & Cellot, B. (1995). On the use of U V spectrophotometry to assess dissolved organic carbon origin variations in the Upper Rhone River. Aquatic Sciences, 57(1), 70–80.
SNIS – Sistema Nacional de Informação sobre Saneamento (2021). Diagnóstico dos serviços de Água e Esgoto − 2020 Brasília: SNIS.
Trujillo-Jiménez, M. A., Liberoff, A. L., Pessacg, N., Pacheco, C., Díaz, L., & Flaherty, S. (2022). SatRed: New classification land use/land cover model based on multi-spectral satellite images and neural networks applied to a semiarid valley of Patagonia. Remote Sensing Applications: Society and Environment, 26(January). https://doi.org/10.1016/j.rsase.2022.100703
Tu, J., & Xia, Z. G. (2008). Examining spatially varying relationships between land use and water quality using geographically weighted regression I: Model design and evaluation. Science of the Total Environment, 407(1), 358–378. https://doi.org/10.1016/j.scitotenv.2008.09.031
USGS – UNITED STATES GEOLOGICAL SURVEY (2020). Earth Explorer. Available from: https://earthexplorer.usgs.gov/. Acessed on: Jan 20th, 2020.
Visco, G., Campanella, L., & Nobili, V. (2005). Organic carbons and TOC in waters: An overview of the international norm for its measurements. Microchemical Journal, 79(1–2), 185–191. https://doi.org/10.1016/j.microc.2004.10.018
Walega, A., Amatya, D. M., Caldwell, P., Marion, D., & Panda, S. (2020). Assessment of storm direct runoff and peak flow rates using improved SCS-CN models for selected forested watersheds in the Southeastern United States. Journal of Hydrology: Regional Studies, 27(August 2019), 100645. https://doi.org/10.1016/j.ejrh.2019.100645
Wang, X., & Zhang, F. (2018). Effects of land use/cover on surface water pollution based on remote sensing and 3D-EEM fluorescence data in the Jinghe Oasis. Scientific Reports, 8(1), 1–13. https://doi.org/10.1038/s41598-018-31265-0
Weishaar, J. L., Fram, M. S., Fujii, R., & Mopper, K. (2003). Evaluation of Specific Ultraviolet Absorbance as an Indicator of the Chemical Composition and Reactivity of Dissolved Organic Carbon. 4702–4708. https://doi.org/10.1021/es030360x
Westerhoff, P., & Anning, D. (2000). Concentrations and characteristics of organic carbon in surface water in Arizona: Influence of urbanization. Journal of Hydrology, 236(3–4), 202–222. https://doi.org/10.1016/S0022-1694(00)00292-4
Wilkinson, K. J., Joz-Roland, A., & Buffle, J. (1997). Different roles of pedogenic fulvic acids and aquagenic biopolymers on colloid aggregation and stability in freshwaters. Limnology and Oceanography, 42, 1714–1724. https://doi.org/10.4319/lo.1997.42.8.1714
Wilson, C. O. (2015). Land use/land cover water quality nexus: quantifying anthropogenic influences on surface water quality. Environmental Monitoring and Assessment, 187(7). https://doi.org/10.1007/s10661-015-4666-4
Zhang, Q., Shen, J., & Sun, F. (2021). Spatiotemporal differentiation of coupling coordination degree between economic development and water environment and its influencing factors using GWR in China’s province. Ecological Modelling, 462(September), 109794. https://doi.org/10.1016/j.ecolmodel.2021.109794
Zhao, W., Zhu, X., Sun, X., Shu, Y., & Li, Y. (2015). Water quality changes in response to urban expansion: spatially varying relations and determinants. Environmental Science and Pollution Research, 22(21), 16997–17011. https://doi.org/10.1007/s11356-015-4795-x
Zumstein, J., & Buffle, J. (1989). Circulation of pedogenic and aquagenic organic matter in an eutrophic lake. Water Research, 23(2), 229–239. https://doi.org/10.1016/0043-1354(89)90047-X

No competing interests reported.

LUPCscript.docx

Water Quality Assessment for Organic Matter Load in Urban Rivers Considering Land Use Dynamics

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

The LUPC model

Materials and methods

Study Area

Laboratory Parameters

Satellite images

The LUPC 1.0 program

LUCP Model Validation

RESULTS AND DISCUSSION

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1