Maximum flow regionalization of the Pomba River basin and flood prevention recommendations

(1)

Institute for Biodiversity and Ecosystem Dynamics

Examiner and daily supervisor: dhr. dr. L.H. (Erik) Cammeraat Co-assessor: dr. ir. E.E. (Emiel) van Loon

MSc. Earth Sciences / Track: Environmental Management Research Thesis 30 EC

João Luca Amorim Tavares e Horta Student Number: 10855637

Maximum flow regionalization of the Pomba River basin and flood

prevention recommendations

(2)

Summary

1. INTRODUCTION ... 6

1.1 Literature Review on flow regionalization ... 8

1.2 General objective ... 9

1.3 Specific objectives ... 10

2. MATERIALS AND METHODS ... 11

2.1 Study Area ... 11

2.2 Diagnosis of Fluviometric Stations ... 13

2.3 Researching the Distribution of Data ... 15

2.4 Maximum Flow Calculation ... 16

2.5 Flow Regionalization ... 20

2.6 Flood Prevention ... 22

2.6.1 Non Structural Measures ... 23

2.6.2 Structural Measures ... 24

3. Results ... 26

3.1 Maximum Flow Calculation ... 26

3.1.1 Flow Continuity Analysis ... 27

3.2 Flow Regionalization ... 30

4. Discussion ... 36

4.1 CPRM Study ... 37

4.2 Flooding and risks... 41

5. Conclusion ... 43

(3)

List of Figures

Figure 1: Location of the Pomba River basin. (Guedes et al., 2009) ... 11

Figure 2: Relief of the Pomba River Basin (CPRM, 2017). ... 12

Figure 3: Record of maximum flow rates in Hidro 1.4. ... 20

Figure 4: The evolution of flood risk management practice (UNESCO, 2013). ... 23

Figure 5: Typical floodwall (Rickard, 2009). ... 25

Figure 6: Seawall in the Maldives (Kapoor, 2020). ... 25

Figure 7: Itaipu Dam and its reservoirs (Itaipu Binacional, 2010). ... 25

Figure 8: Flow regionalization curve Q Rt 10 years. ... 32

Figure 9: Flow regionalization curve Q Rt 25 years. ... 33

Figure 10 Flow regionalization curve Q Rt 50 years. ... 33

Figure 11: Flow regionalization curve Q Rt 100 years... 34

Figure 12: Reported flood phenomena between 1980 and 2010 (European Environment Agency, 2015). ... 42

(4)

List of Tables

Table 1: Relief classification (CPRM, 2017). ... 13

Table 2: Fluviometric stations initially able to be the object of the study. ... 14

Table 3: Selected fluviometric stations. ... 15

Table 4: Estimation of flows in the recurrence times of 10, 25, 50 and 100 years. ... 26

Table 5: Flow continuity analysis (Q Rt 10 years). ... 28

Table 9: Flow continuity analysis in the common period. ... 30

Table 10: Selected stations for the study of flow regionalization. ... 31

Table 11: Regional equations and coefficients of determination. ... 34

Table 12: Dimensionless values of the maximum annual flows Q / Qmc, to its return period RT (CPRM, 2003). ... 38

Table 13: Maximum flow rates determined by the CPRM equation. ... 39

Table 14: Comparison of results between regional equations of CPRM and this dissertation. ... 40

(5)

ABSTRACT

The Pomba River basin is the largest tributary of the Paraíba do Sul River, the richest water region in Brazil, but studies of maximum flows in this tributary are rare, more focused on minimum and medium flows. These studies are important to understand the hydrological behavior of the basin, but the country suffers from a monitoring network below what is necessary. With this difficulty, hydrologists have developed techniques to transfer data from one location to another, in the same basin, called Flow Regionalization, which after a direct linear regression generates an equation capable of estimating flow rates in locations where there are no fluviometric stations. In this dissertation thesis, floods and their consequences will be discussed. In addition, the maximum flows will be calculated and regionalized in the recurrence times of 10, 25, 50 and 100 years. These flows are used as flows for projects in microdrainage and macrodrainage. The results showed that the study had a satisfactory outcome, since the lowest correlation coefficient was 0.9368, so the regional equations can be used in non-structural measures to combat floods, such as the elaboration of flood maps. Finally, a comparison of results with flow sources is made using the CPRM regional equation, the last major study of maximum flows and flow regionalization in the Pomba River basin, published in 2003, in order to evaluate the results and discuss the possible differences between them.

(6)

6

1. INTRODUCTION

Brazil is the country with the largest reserve of superficial fresh water in the world, excluding polar caps, with about 12% of all available water (Geo Water Resource Brazil, 2007). However, the water distribution in the national territory is unbalanced, concentrating most of the surface water in the northern region of the country, which is one of the least economically developed and the least populated (Geo Water Resource Brazil, 2007).

In water resource management, it is important to know the water availability of an area and the risks involved in water flows. This study aims to study the maximum flows of a sample basin in Brazil.

The southeastern region of Brazil is richer and has the largest population (IBGE, 2010), making it essential to know its hydrological behavior in order to efficiently manage water.

The management of water resources appears as one of the most important tools of a society, since it is possible to make water consumption compatible for different uses, such as human supply, agriculture and industry.

In this sense, it is important to create legislation to maximize its use. The Brazilian National Water Resources Law No. 9433 of 8 January 1997 was enacted to regulate its management, which defined the multiple management of water resources, created the National Water Resources Policy and the Water Resources Information System. It also determined that water is in the public domain and endowed it with economic value and that in cases of water scarcity the priority uses will be for human consumption and animal hydration. Furthermore, basins were defined as the territorial unit for the implementation of the National Water Resources Policy and performance of the National Water Resources Management System (Brazilian Federal Government, 1997).

According to Tucci et al. (2007), a hydrographic basin is an area with a set of slopes and a drainage network, draining all water bodies and precipitation towards a single point of eluvium.

With the above, an adequate knowledge of the hydrological behavior of hydrographic basins becomes essential for the efficient management of water resources. Minimum and average flows can be applied to water supply management and energy generation, whilst maximum flows can be applied to flood-preventing

(7)

7

hydraulic works (eg. micro drainage of dams). Fluvial and pluvial monitoring networks are essential for this purpose, in addition to knowledge of the physical and climatic characteristics of the basin (Tucci et al., 2007).

However, Brazil has considerably reduced its river and rainfall monitoring networks in its basins, due to recurring economic crises, implying a lack of investment in new stations and a lack of financial resources to maintain the existing ones, a situation aggravated in small basins.

To face this problem and maximize available information hydrologists have developed a technique capable of transferring hydrological information from one location to another belonging to the same basin, known as Flow Regionalization (Tucci et al., 2007). The author explains that this procedure seeks to the maximum use of the existing information in order to estimate hydrological parameters in places that do not have data or low quality data.

The Paraíba do Sul River basin, which the Pomba River basin is part of, is one within the largest number of hydrological studies in Brazil, several studies of flow determination and flow regionalization. This is due to its relevance of guaranteeing the availability of water in the richest region of the country. However, studies of extreme flows are rarer, more focused on dams.

The United Nations (2001) estimates that floods are responsible for one third of the cost resulting from natural disasters and this represents two thirds of the affected population at the beginning of the century.

Climate change is contributing to the increase in the registration of floods, making them more frequent and of greater magnitude (IPCC, 2012). This leads to the need to conduct more and more studies of extreme flows.

The last public study on maximum flows and flow regionalization in the Pomba River basin was carried out by the Geological Survey of Brazil in 2003 (CPRM, 2003). Therefore, there is a need to conduct further studies in the region to update the hydrological knowledge of the Pomba River basin.

The present project will address the determination of the maximum flows in the Pomba River basin, a tributary of the Paraíba do Sul River, and will carry out the Regionalization of Flow of the basin, in order to provide support for the water management of the region.

(8)

8

Finally, a comparison will be made of the application of the regional equations determined by CPRM in the 2003 study with those produced in this thesis, in order to identify the differences between them.

In the literature, studies of regionalization of maximum flows are less frequent than of average and minimum flows, meaning a scarce international theoretical framework. With this gap, this thesis intends to contribute with a technical study in this area.

1.1 Literature Review on flow regionalization

Flow regionalization approaches have been carried out in many different systems, especially minimum flow regionalization. Such studies were aimed at estimating the water availability of a region and guaranteeing water supply, transport, ecological sustainability and hydroelectrical use.

Laaha and Blöschl (2006) carried out a broad project comparing 4 different types of Q95 flow regionalization, reference flow recorded in 95% of the observation period. The study covered 325 sub-catchments in Austria and found that the grouping of regions based on seasonality regions performs best.

Mamum et al. (2009) used a technique of multiple variables to determine the magnitude and frequency of 1, 7 and 30-day low flows on the peninsula of Malaysia, where similar low flow frequency curves were grouped together.

Haberlandt et al. (2001) investigates possibilities for the regionalization of flow components within large river basins, with the river Elbe as a case study. The study concludes that the average base flow index is strongly related to topographical, pedological, hydrogeological and precipitation characteristics.

Although it is common to find studies of regionalization of minimum and average flows in the international literature, the same cannot be said about maximum flows.

Lee et al. (2005) demonstrated that the most suitable types of precipitation models to be used in the United Kingdom were used to determine extreme flows. Viviroli et al. (2009) also used rainfall-flow modeling, more specifically of continuous-flow precipitation, developed in Switzerland. For this purpose, 140 hydrographic basins were calibrated, generating a regionalization of model parameters.

(9)

9

Coronado-Hernández et al. (2020) used 362 rainfall stations in Colombia to identify the best hydrological probability distribution for extreme rain events. The study estimated flows for the return periods of 5, 10, 25, 50 and 100 years, but did not perform the flow regionalization.

Odry and Arnaud (2017) applied an extensive study in 1535 French catchments comparing methods of flood frequency analysis from statistics to different types of regionalization. They suggest that flood frequency analysis from the same statistical family largely relies on the regionalization step and emphasizes the difficulty of the regionalization process.

Lopes et al. (2017), Matos et al. (2020), Albuquerque et al. (2020) and Tucci (1995) conducted studies of regionalization of maximum flows in Brazilian hydrographic basins, but none of them addressed the Paraíba do Sul river basin, the most economically important. Therefore, this thesis will conduct a study of regionalization of maximum flows in the Pomba River basin, a sub-basin of the Paraíba da Sul River, in order to cover a research gap in one of the most important regions of Brazil.

1.2 General objective

This study aims to discuss methods and their suitability for determining maximum flows and regionalization of flow. In addition, to calculate the maximum flows in the recurrence times of 10, 25, 50 and 100 years of the Pomba River basin. With the results, studies of flow regionalization will be made in each recurrence time. Finally, a comparison will be made of the use of the regional equations found with the regional equations of the Mineral Resources Research Company (CPRM).

(10)

10

1.3 Specific objectives

• Discuss floods, their impacts, and ways of preventing flooding;

• Discuss methods for determining maximum flows and regionalization of flow;

• Diagnosis of the hydrological network of the Pomba River basin;

• Determine the maximum flows in the recurrence times of 10, 25, 50 and 100 years;

• Determine the regionalization curve and equations;

• Compare the regional equations obtained with those from the CPRM study (2003);

• Discuss floods and their risks;

(11)

11

2. MATERIALS AND METHODS 2.1 Study Area

The watershed of the Pomba River is the study area of this thesis. The Pomba River, with its source in the municipality of Barbacena, Minas Gerais, at 1,100 meters of altitude and its mouth on the Paraíba do Sul River, after covering 265 km of its main river bed. The watershed of the Pomba River is located between the coordinates 43°45’15,3" and 41°59'2,5" West (longitude); and 20°51'58,0" and 21°42'53" South (latitude), covering three Brazilian states: São Paulo, Minas Gerais and Rio de Janeiro.

Figure 1: Location of the Pomba River basin. (Guedes et al., 2009)

According to the Pomba River Action Book (2006), the basin has a drainage area of 8,616 km² and an estimated population of 450,000 inhabitants. Plus, it includes 35 municipalities in Minas Gerais, 3 municipalities in Rio de Janeiro and the main tributaries of the rivers Novo, Xopotó, Formoso and Pardo.

(12)

12

The Pomba River Action Book (2006) also shows that environmental sanitation in the basin is precarious, with a large part of domestic sewage being released in natura directly into the water courses of the basin. As for forest cover, there is a high degree of deforestation, mainly in the headwaters, with a total absence of primary and secondary vegetation cover, causing relevant erosion along the margins of water courses. These sediments result in the relevant silting up of rivers, impairing their quality.

Deforestation in the Pomba River basin started in the coffee cycle in Brazil, during the 19th and early 20th centuries, with land being severely deforested, even in regions of river springs, contributing to the worsening of water quality, adding to the release of industrial and domestic effluents.

The Pomba River basin has an Aw climatic, hot humid tropical climate, with a typical rainy summer and dry winter climate (Köeppen, 1948). The basin has these two well-defined seasons: rainy, from October to March, and dry, from May to September.

The basin relief is divided into 6 classes: mountainous (32.1%), smooth wavy (29.6%), strong wavy (17.2%), flat (10.2%), wavy (8.9% ) and mountainous fort (2%) (CPRM, 2017).

(13)

13 Table 1: Relief classification (CPRM, 2017).

Declivity (%) Relief type % of basin area

0 - 3 Flat 10,2 % 3 - 8 Smooth wavy 29,6 % 8 - 20 Wavy 8,9 % 20 - 45 Strong wavy 17,2 % 45 - 75 Mountainous 32,1 % >75 Mountainous fort 2,0 %

According to the exploratory soil map related to the Paraíba do Sul basin, carried out by the Brazilian Agricultural Research Corporation (EMBRAPA), the Pomba River basin has the following soil types: Cambisols, Ferrasols and Acrisols (EMBRAPA, 2011).

Cambisols have a certain degree of evolution, but still have primary minerals, they occur in mountainous reliefs, strong undulating and undulating. Red - Yellow Latosols are well-drained to markedly well-drained soils, with moderate to imperfect drainage in flat / undulating reliefs. The Acrisols may or may not have gravels, but probably have an argic horizon, reducing infiltration at the B horizon, admitting wide variability of textural classes, where stones may be occasionally present (CPRM, 2017).

2.2 Diagnosis of Fluviometric Stations

The database used in this thesis is on the hidroweb website, which belongs to the Brazilian National Water Agency, responsible for managing the hydrometeorological information of the national monitoring network, one of the obligations imposed by federal law No. 9433 (Brazilian Federal Government, 1997).

According to the survey carried out on the hidroweb website, 102 fluviometric stations were listed belonging to the Pomba River basin. However, only 24 stations had a flow record history, and it is not possible to take advantage of the other 78 stations for not having the desired flow data or that did not have any data at all. Table 2 shows the fluviometric stations object of this study.

(14)

14 Table 2: Fluviometric stations initially able to be the object of the study.

Code Name River Drainage

Area (km²)

State Municipality

58710000 Usina Itueré Pomba 784 Minas Gerais Rio Pomba 58720000 Tabuleiro Formoso 322 Minas Gerais Tabuleiro 58725000 Fazenda Ferraz Formoso 387 Minas Gerais Rio Pomba 58730000 Guarani RV Pomba 1660 Minas Gerais Guarani 58730001 Guarani Pomba 1650 Minas Gerais Guarani 58731000 PCH Ivan Botelho II

Jusante

Pomba 1790 Minas Gerais Guarani 58731300 PCH Zé Tunin

Montante

Pomba 1810 Minas Gerais Guarani 58731500 PCH Zé Tunin

Barramento

Pomba 305 Minas Gerais Guarani 58735000 Astolfo Dutra Pomba 2350 Minas Gerais Astolfo Dutra 58736000 Barra do Xopotó Xopotó 1280 Minas Gerais Astolfo Dutra 58737080 PCH Ervália

Montante desativada

Bagre 54.5 Minas Gerais Ervália

58737180 PCH Ervália Jusante Bagre 73 Minas Gerais Guiricema

58750000 Piau Piau 490 Minas Gerais Piau

58753080 PCH Anna Maria Montante

Pinho 95.7 Minas Gerais Santos Dumont 58755000 Rio Novo Novo 835 Minas Gerais Rio Novo 58765000 Usina Maurício Novo 1910 Minas Gerais Itamarati de

Minas 58765001 Usina Maurício Novo 1770 Minas Gerais Leopoldina 58770000 Cataguases Pomba 5880 Minas Gerais Cataguases 58788050 Vale do Pomba Pomba 6850 Minas Gerais Leopoldina 58790000 Santo Antônio de

Pádua

Pomba 8210 Rio de Janeiro Santo Antônio de Pádua

58790002 Santo Antônio de Pádua II

For a consistent hydrological study, it is desirable to obtain data covering a minimum period of 10 consecutive years of observation (CPRM, 2017). In the Pomba River basin it was not possible to use 7 stations because they did not meet this criteria. In addition, relocated stations were excluded, since they had an incongruent flow history, which would be detrimental to the linear regression that will be made for the regionalization of flow rate. Thus, the ones with the longest observation period were selected.

(15)

15

As the hydrological monitoring network in the basin is small, it was not possible to work with only the same temporal flow records from fluviometric stations. As a result, as much data as possible was used. Table 3 shows the selected fluviometric stations.

Table 3: Selected fluviometric stations.

Area (km²)

State Municipality Observation Period (Years) 58710000 Usina Itueré Pomba 784 Minas

Gerais

Rio Pomba 1929 - 2014 58720000 Tabuleiro Formoso 322 Minas

Gerais

Tabuleiro 1943 - 2014 58725000 Fazenda Ferraz Formoso 387 Minas

Gerais

Rio Pomba 1930 - 1964 58730001 Guarani Pomba 1650 Minas

Gerais

Guarani 1949 - 2014 58735000 Astolfo Dutra Pomba 2350 Minas

Gerais Astolfo Dutra 1931 - 2014 58736000 Barra do Xopotó Xopotó 1280 Minas Gerais Astolfo Dutra 1988 - 2014

58750000 Piau Piau 490 Minas

Gerais

Piau 1931 - 2014 58755000 Rio Novo Novo 835 Minas

Gerais

Rio Novo 1943 - 2014 58765001 Usina Maurício Novo 1770 Minas

Gerais

Leopoldina 1963 - 2014 58770000 Cataguases Pomba 5880 Minas

Gerais Cataguases 1930 - 2014 58790000 Santo Antônio de Pádua Pomba 8210 Rio de Janeiro Santo Antônio de Pádua 1935 - 2002

2.3 Researching the Distribution of Data

An asymmetry coefficient (g) is used to measure the degree of dispersion of a data distribution; it is used as a variable for different types of maximum flow calculation methods.

Tucci et al. (2007) shows a negative asymmetry coefficient indicating the sample's median is greater than the average, while positive asymmetry indicates that the sample's average is greater than the median. Null asymmetry coefficient indicates that the mean and median are equal. Therefore, the asymmetry coefficient indicates the degree of dispersion of the sample.

(16)

16

The Brazilian Water Agency (2010) addresses that statistical studies preferably use the Exponential Distribution of Two Parameters when the asymmetry coefficient is greater than 1.5, and the Gumbel Distribution when the asymmetry coefficient equal to or less than 1.5. The asymmetry coefficient (g) is calculated by Equation 1. 𝑔 = 𝑛 (𝑛 − 1)(𝑛 − 2)( ∑𝑛 (𝑥_𝑖 − 𝑥)3 𝑖=1 𝑠3 ) (1) Where, • n: sample quantity;

• 𝑥𝑖: sample value in position i;

• 𝑥: arithmetic mean of the sample; • 𝑠: standard deviation of the sample.

𝑠 = √∑ (𝑥𝑖 − 𝑥)

2 𝑛

𝑖=1

𝑛 − 1 (2)

2.4 Maximum Flow Calculation

A maximum flow in an average recurrence time T is expected to be equaled or exceeded in a certain river section. For the flow estimation there are basically two methods: hydrometeorological and statistical. The first method is based on the premise of the highest possible precipitation, estimating the maximum probable precipitation. Statistical methods, on the other hand, are worked on by analyzing hydrological information on the spot or in regions of homogeneous behavior. There are different statistical methods, depending on the quantity and quality of data in the region (Scuissiato, 2013).

Some of the most common statistical methods for determining maximum flow rates will now be exemplified:

(17)

17

For Log-Normal Distribution Type II we have Equation 3 (Chow, 1951).

𝑄 = ⌊𝑒 √ln(𝑧2₊₁₎ 𝑇 − ln(𝑧2+1) 2 ⌋ − 1 𝑧 (3) Where,

• Q: maximum flow at recurrence time T; • T: return period;

• 𝑧 = 𝜎

𝜇;

• 𝜎: sample standard deviation; • 𝜇: sample mean.

For Log-Normal Distribution Type III we have Equation 4 (Bobée; Robitaille, 1977). 𝑄 = ⌊𝑒 √ln(𝑧2₊₁₎ 𝑇 − ln(𝑧2+1) 2 ⌋ − 1 𝑤 (4) Where,

• Q: maximum flow at recurrence time T; • T: return period;

• 𝑧 = 𝜎

𝜇;

• 𝜎: sample standard deviation; • 𝜇: sample mean. 𝑤 = 1 − 𝜔 3 2 𝜔13 (5) 𝜔 = −𝛾 + √𝛾 2_{+ 4} 2 (6) Where,

(18)

18

For Pearson Type III Distribution or Type III Gamma Distribution we have Equation 7 (Bobée; Ashkar,1991).

𝑄 = 𝐷 + (𝐷2− 1)𝛾 6+ 1 3(𝐷 − 6𝐷) ( 𝛾 6) 2 − (𝐷2− 1) (𝛾 6) 3 + 𝐷 (𝛾 6) 4 +1 3( 𝛾 6) 5 (7) 𝐷 = 𝑇 − 2,30753 + 0,27061𝑇 1 + 0,99229𝑇 + 0,04481𝑇2(8) Where,

• Q: maximum flow at recurrence time T; • 𝛾: asymmetry coefficient g (Equation 1); • T: return period.

For the Exponential Distribution of Two Parameters, Equation 9 is used to estimate the maximum flow in a given recurrence time (Eletrobrás, 2020).

𝑥_𝑇 = 𝑥₀− 𝛽 ln (1 𝑇) (9) Where,

• 𝑥_𝑇: maximum flow at recurrence time T; • T: return period;

• 𝑥0 = 𝑥 − 𝑠;

• 𝛽 = 𝑠;

(19)

19

For Gumbel Distribution, Equation 10 is used to estimate maximum flow in a given recurrence time (Eletrobrás, 2020).

𝑥_𝑇 = 𝜇 − 𝛼 (ln (− ln (1 − 1

𝑇))) (10)

Where,

• 𝛼 = 0,78𝑠;

• 𝑠: standard deviation of the sample; • 𝜇 = 𝑥 − 0,577𝛼;

• 𝑥: arithmetic mean of the sample; • T: return period.

In this study, only the Gumbel and Exponential Distribution of Two Parameters will be used, as they are widely used in Brazil and by its national water agency.

With the series of average daily flow rates, the highest value in each year is selected. From the established series of annual maximums, the average, standard deviation and asymmetry are calculated. From the analysis of the asymmetry value, the statistical distribution is chosen (Gumbel or Exponential), and the project flows are defined.

The National Water Agency - ANA makes the Hidro 1.4 software available for free download on its website (in the Reference List). The software makes it possible to view the various data from fluviometric stations, such as flow record, average flow rates, quotas, among other information.

To calculate the maximum flows, data from the selected fluviometric stations in the Pomba River basin were downloaded after the diagnosis made in the section 2.2.

Hidro 1.4 also provides information on maximum monthly flow rates observed (from its series of average daily flow rates), from which the highest value registered annually is identified (Figure 3). These data were extracted into an excel spreadsheet, where the mean, standard deviation and asymmetry coefficient were calculated, and according to the value of this coefficient, the Gumbel or Exponential Distribution of Two Parameters is determined, and then maximum flow rates are calculated with the recurrence times of 10, 25, 50 and 100 years.

(20)

20 Figure 3: Record of maximum flow rates in Hidro 1.4.

It is worth noting that the data used in this study was formatted by ANA and no raw data was analyzed. In the formatted data there are some with special definitions, namely: estimated (*), doubtful (?) or dry measuring rule (#). In the stations selected in this thesis, some presented data considered estimated (*), however, after comparing with data from stations belonging to the same river or sub-basin, it was verified that these data could continue as the object of this study.

2.5 Flow Regionalization

Several authors suggest that after determining the maximum flow rates, through the most appropriate statistical distribution, they should have their flow rates dimensioned (Tucci, 2002; Castellarin, 2007; Naghettini, 2017).

(21)

21

This generates dimensionless curves to determine homologous regions hydrologically, where there is the same probability distribution. Then, regional curves and equations are generated for each homogeneous region through direct linear regression (Tucci, 2002).

Due to the scarcity of data in the location / basin of interest, it is sometimes chosen to adopt a regional curve that covers the extreme values, calculated in surrounding basins or in stations located in the same basin, and to transfer, from that curve, the extreme flow values for the study site (Eletrobrás, 2020).

From the calculation of recurrent maximum flows of the existing fluviometric stations, the regression curves of these variables are determined, having the respective drainage areas as a dependent variable. The curves are defined by Equation 11.

𝑄_𝑡 = 𝑎(𝐴)𝑏(11)

Where,

• a and b: coefficients;

• 𝑄_𝑡: specific flow rate for the recurrence time (T), in l/s.km²; • A: drainage area of each location / post, in km².

In some studies, other physical and climatic characteristics of the basin are used as variables in the regression equation, in addition to the drainage area, such as slope and average river length, drainage density and average annual precipitation. In this, only the drainage area was used as an independent variable due to the lack of data on the other variables.

To verify the quality of the linear correlation calculated between the variables, the R² correlation coefficient is evaluated. The analysis of this coefficient indicates the degree of adjustment between the flow and the drainage area, dependent and independent variable, respectively. The R² value closer to 1 indicates a high degree of adjustment between the points with the determined regression curve (Eletrobrás, 2020).

Furthermore, the Barra de Xopotó station was excluded from this study because of a flow regionalization analysis made on the Paraíba do Sul river basin by

(22)

22

the Geological Survey of Brazil - CPRM (CPRM, 2017). In this study the homogeneous regions were determined, that is, regions with several stations with series coming from populations oriented by the same probability distribution, varying parameters between them. The Barra de Xopotó station was then highlighted as an exceptional region due to its data heterogeneity.

2.6 Flood Prevention

Floods are natural phenomena that can occur more frequently or rarely, depending on the location. They depend on the intensity of the rain, shape of the basin, use and occupation of the soil and degree of impermeabilization, among other factors (Tucci et al., 2007).

Due to disorderly growth in developing countries like Brazil, floods have become more and more frequent. Some of the main reasons are: occupation on the banks of river beds and increased impermeabilization of the soil and river beds, transforming them into urban channels (Tucci et al., 2007).

According to the European Commission (2003), mitigation and non-structural measures are more sustainable long-term solutions for the management of water resources. However, structural measures are still important and must be well dimensioned.

Unesco (2013) points out that floods were important for the development of civilizations and that humanity learned to deal with its effects. From the 1960s onwards, large flood control structures emerged, but it was realized that only engineering was not enough and the concept of risk in decision-making should be used.

(23)

23 Figure 4: The evolution of flood risk management practice (UNESCO, 2013).

Non Structural Measures are any measures not involving physical construction to reduce and prevent risks and impacts, using policies and laws, for example (Climate Service Center, 2013). Some structural measures are floodwalls / seawalls, dam and reservoirs (CRUE, 2008).

2.6.1 Non Structural Measures

The urbanization of Brazilian cities occurred in a disorganized manner, due to the country's rapid economic growth around the middle of the last century, associated with the increase in the population growth rate (IBGE, 2010). Such factors resulted in irregular occupations in several areas that should have been protected, such as hillsides and marginal strips of rivers.

An effective non-structural measure for flood prevention is the relocation of populations that occupy marginal areas of rivers to safe locations, in addition to developing programs to raise awareness of the risks of occupying flood plains.

In addition, areas of springs and marginal strips of rivers should be reforested, thus reducing erosion of the banks and silting up the river channels, in addition to increasing the infiltration of water into the soil (Carmo, 2015).

The results of this thesis will also be useful for the planning of non-structural measures, since with the estimation of flows in the recurrence times of 10, 25, 50 and

(24)

24

100 years, it will be possible to elaborate flood risk maps of the areas potentially affected for each recurrence period.

These flood risk maps will allow decision makers to assess in advance the risks inherent in each area according to the flood flow in each recurrence period, elaborating specific action plans for each hydrological event.

Still as a non-structural measure, CPRM (2020) developed the project Hydrological Alert System of the Pomba River Basin, which aims to monitor the rainfall regime in the basin and to monitor the increase in river quotas, pre- determining levels of flood risk along the basin. This system allows for the prediction of extreme event situations, enabling preventive and mitigating actions by the competent bodies.

2.6.2 Structural Measures

In Brazil it is common to use dams as hydraulic structures for various uses, such as power generation, since energy from hydroelectric sources is the main source of electricity in the country, representing 61.1% (MME, 2020). These structures are also widely used for irrigation, water supply, flood control, among other uses (Sória, 2008).

A dam is defined as a structure built perpendicular to a river or thalweg, aiming at increasing the water level and / or creating a reservoir of water to regulate flow (MIN, 2002).

Barth (1997) points out that the flood control reservoirs are important to manage the flow drained downstream, in addition to being an important mechanism for attenuating the flood peaks. These structures are typically composed of an input structure, energy dissipation and water release.

Pomba River Action Book (2006) highlights that the basin has less severe floods than the neighboring Muriaé sub-basin, also a tributary of the Paraíba do Sul River, probably due to the contribution of the existing reservoirs along the Pomba River basin.

(25)

25 Figure 5: Typical floodwall (Rickard, 2009).

Figure 6: Seawall in the Maldives (Kapoor, 2020).

(26)

26

3. Results

3.1 Maximum Flow Calculation

After transferring the data to a spreadsheet and calculating the asymmetry coefficients of the stations, and thus determining the statistical distribution to be used, it was possible to estimate the maximum flow in each of them (Table 4).

Table 3: Estimation of flows in the recurrence times of 10, 25, 50 and 100 years.

Code Station Asymmetry

Coefficient Statistical Model Drainage Area (km²) Q Rt 10 years (m³/s) Q Rt 25 years (m³/s) Q Rt 50 years (m³/s) Q Rt 100 years (m³/s)

58710000 Usina Itueré 0,51666467 Gumbel 784 176 212 239 266

58720000 Tabuleiro 1,770115058

Exponential of two parameters

322 109 140 164 188

58725000 Fazenda Ferraz 0,059578814 Gumbel 387 54 61 67 72

58730001 Guarani 0,55388666 Gumbel 1650 285 336 374 412

58735000 Astolfo Dutra 0,68900598 Gumbel 2350 352 415 461 507

58750000 Piau 0,430566531 Gumbel 490 88 105 118 131

58755000 Rio Novo 1,081390767 Gumbel 835 146 174 195 215

58765001 Usina Maurício 1,236926451 Gumbel 1770 345 416 468 520

(27)

27 58790000 Santo Antônio de

Pádua 0,757058877 Gumbel 8210 1433 1710 1915 2119

3.1.1 Flow Continuity Analysis

Flow continuity analysis is a tool to verify a hydrological premise, as it is expected that the characteristic flow of a section of a river will increase as the drainage area increases, that is, a downstream station must necessarily have a flow greater than an upstream station. If a flow from a downstream station is verified to be lower than the flow from an upstream station, it is said that there is a negative flow increase. In such cases it is advisable to check the use of water in the analyzed section (Tucci et al., 2007).

Thus, the premise of the flow continuity analysis of the Pomba River basin was made:

• Q Fazenda Ferraz > Q Tabuleiro

• Q Guarani > Q Usina Itueré + Q Fazenda Ferraz • Q Astolfo Dutra > Q Guarani

• Q Rio Novo > Q Piau

• Q Usina Maurício > Q Rio Novo

• Q Cataguases > Q Usina Maurício + Q Astolfo Dutra • Q Santo Antônio de Pádua > Q Cataguases

In order to verify this premise, the continuity of flows in each recurrence time was analyzed (Tables 5, 6, 7 and 8).

(28)

28 Table 4: Flow continuity analysis (Q Rt 10 years).

Q Rt 10 years (m³/s)

Fazenda Ferraz ≥ Tabuleiro Continuity of flows

54 109 Wrong

Guarani ≥ Fazenda Ferraz + Usina Itueré

285 163 Ok

Astolfo Dutra ≥ Guarani

352 285 Ok

Rio Novo ≥ Piau

146 88 Ok

Usina Maurício ≥ Rio Novo

345 146 Ok

Cataguases ≥ Usina Maurício + Astolfo Dutra 835 697 Ok Santo Antônio de Pádua ≥ Cataguases 1433 835 Ok

Table 5: Flow continuity analysis (Q Rt 25 years). Q Rt 25 years (m³/s)

61 140 Wrong

336 201 Ok

415 336 Ok

Rio Novo ≥ Piau

174 105 Ok

416 174 Ok

(29)

29 Table 6: Flow continuity analysis (Q Rt 50 years).

Q Rt 50 years (m³/s)

67 164 Wrong

374 231 Ok

461 374 Ok

Rio Novo ≥ Piau

195 118 Ok

468 195 Ok

Table 7: Flow continuity analysis (Q Rt 100 years). Q Rt 100 years (m³/s)

72 188 Wrong

412 260 Ok

507 412 Ok

Rio Novo ≥ Piau

215 131 Ok

520 215 Ok

(30)

30

Analyzing the Tables 5, 6, 7 and 8, we verify that the flow continuity between Fazenda Ferraz and Tabuleiro stations is not met. Then an analysis was made of their extreme flows in a period of common observation between them, totaling an equal 11 years of flow recordings (Table 9).

Table 8: Flow continuity analysis in the common period.

Code Station Asymmetry

Coefficient Statistical Model Drainage Area (km²) Q Rt 10 years (m³/s) Q Rt 25 years (m³/s) Q Rt 50 years (m³/s) Q Rt 100 years (m³/s) 58720000 Tabuleiro 0,464333515 Gumbel 322 45 50 53 57

58725000 Fazenda Ferraz 0,065049191 Gumbel 387 53 60 65 70

Table 9 shows that the flow rates at each recurrence time at the Fazenda Ferraz station were higher than those at Tabuleiro station, thus fulfilling the flow continuity.

It is important to note that in the analysis of the common period the Tabuleiro station was calculated by the Gumbel distribution, as it presented a lower coefficient of asymmetry. With that, it is verified that its flow rates present great variability in the general observation period.

The Fazenda Ferraz and Tabuleiro stations present a continuity of flows analyzing the common period. However, this condition is not met in the general period, so the first was excluded from this study because it has a shorter period of observation.

3.2 Flow Regionalization

After the analysis carried out in this study, of the 102 fluviometric stations listed initially in hidroweb, it was only possible to use 9 of them for the study of flow regionalization, verifying the low quality of the hydrological network in the basin. Table 10 summarizes the stations used for the study of flow regionalization.

(31)

31 Table 9: Selected stations for the study of flow regionalization.

Area (km²)

State Municipality

58710000 Usina Itueré Pomba 784 Minas Gerais Rio Pomba 58720000 Tabuleiro Formoso 322 Minas Gerais Tabuleiro 58730001 Guarani Pomba 1650 Minas Gerais Guarani 58735000 Astolfo Dutra Pomba 2350 Minas Gerais Astolfo Dutra

58750000 Piau Piau 490 Minas Gerais Piau

58755000 Rio Novo Novo 835 Minas Gerais Rio Novo 58765001 Usina Maurício Novo 1770 Minas Gerais Leopoldina 58770000 Cataguases Pomba 5880 Minas Gerais Cataguases 58790000 Santo Antônio de

Pádua

In the literature, studies can be found that use different physical and meteorological characteristics of the basin to determine the regional equation, such as drainage area, average slope, length of the thalweg, precipitation, among others. However, in this study it was decided to use only the drainage area as an independent variable, as it is relatively easy to obtain in geoprocessing software.

In a spreadsheet the data of the maximum flows generated for the recurrence times of 10, 25, 50 and 100 years for each fluviometric station, were plotted in axis in a graph, and in the other axis their respective drainage areas. Afterwards, a potential trend line was generated for each graph in the respective recurrence time, determining its equation and the correlation coefficient R².

To understand the importance of flow estimates at different times of recurrence, a knowledge of probability is necessary. A given recurrence time represents a flow value associated with a risk of being equaled or exceeded, as represented by Equation 12 (Tucci, 2007).

(32)

32 𝑅𝑡 =1 𝑃(12) Where, • Rt: recurrence time; • P: probability.

Thus, flows in the recurrence times 10, 25, 50 and 100 years have a probability / risk of happening given a random year of 10 %, 4 %, 2 % and 1 %, respectively.

(33)

33 Figure 9: Flow regionalization curve Q Rt 25 years.

(34)

34 Figure 11: Flow regionalization curve Q Rt 100 years.

Table 10: Regional equations and correlation coefficient.

Flow (m³/s) Equation R²

Q Rt 10 years Q = 0,6205A8366 _0,9597

Q Rt 25 years Q = 0,8374A8207 _0,9491

Q Rt 50 years Q = 1,005A8119 _0,9424

Q Rt 100 years Q = 1,1746A8048 _0,9368

Analyzing the coefficient of determination R² of the regional equations, all were above 0.9000, indicating good adherence of the flow rates of the fluviometric stations to the regionalization curves. It was identified that with the increase in the recurrence time, the coefficient of determination decreased, indicating a greater variability.

However, all coefficients of determination were greater than 0.9300, indicating that the fluviometric stations in the Pomba River basin are, in fact, located in the same homogeneously hydrological region. These values of correlation coefficients indicate that the data are statistically significant for the basin, if they were not, they would have R² below 0.9000.

The coefficient of determination represents an assessment of the reliability of the linear regression performed and, consequently, of the regionalization of flow.

(35)

35

Therefore, the study of this thesis reached a good reliability assessment, since the lowest correlation coefficient was 0.9368.

In this study, fluviometric stations with drainage areas ranging from 322 to 8,210 km² were used, so the use of regional equations in the Pomba River basin is limited to other locations with drainage areas contained in this interval, as evidenced by Naghettini (2017).

(36)

36

4. Discussion

Matos (2020) et al. regionalized the maximum flows of the Jurema River basin, Brazil. The study concluded that the best results were achieved using drainage area, perimeter and total length of water courses as explanatory variables. These studies showed R² correlation coefficients varied between 0.98 and 0.99 for different recurrence times, using linear and potential regression equations, presenting the same R² value for some recurrence times.

Lopes et al. (2017) regionalized the maximum flows of the Teles Pires basin, Brazil. The study used precipitation of rainiest semester, the total annual precipitation, drainage area, drainage density and length of main water body as explanatory variables. The coefficients of determination R² varied between 0.97 and 0.99 for different times of recurrence, using regression equations of the linear and potential type, with the first presenting higher R².

Both studies showed equal or close coefficients of determination using the drainage area in the regression in comparison with the other explanatory variables. Therefore, a regionalization of flow with the drainage area as a variable is sufficient in most cases, giving greater reliability in the results found in this thesis.

Tucci et al. (2007) suggests that regionalization of flows with a correlation coefficient far from 1 represents an area of great variability, so the fluviometric stations should be regrouped in order to compress homogeneous regions. Therefore, flow regionalizations must be carried out for each of the regions. In this thesis, this procedure was not necessary, since the lowest correlation coefficient was 0.9368.

Euclydes et al. (2001) made the regionalization of the minimum, average and maximum flows in the upper São Francisco basin. For maximum flows, the study determined three hydrologically homogeneous regions, two of which showed better results using only the drainage area as an explanatory variable, while the other performed better with the drainage area and the slope of the main river.

For the 50-year return period, the authors found a coefficient of determination of 0.83 and 0.98 for the regions that used only the drainage area. The region that used the drainage area and the slope of the main river obtained an R² of 0.95.

In order to have a more accurate comparison, it is necessary to study the same basin, which is why the results of this thesis were related to the flow regionalization carried out by CPRM in 2003. But it is important to emphasize that the

(37)

37

results of this thesis are comparable in statistical reliability to other studies carried out. Comparing with more studies in the same basin would be good, but it is not feasible at the moment.

For future work, further studies of other methodologies are recommended, in order to further increase the value of the coefficient of determination. Another alternative would be to increase the number of explanatory variables, in addition to the drainage area. For example, use the average rainfall variable for the wettest period, drainage density and length of main water body and see if this increases the correlation coefficient.

4.1 CPRM Study

In 2003, the Geological Survey of Brazil carried out the last major study of regionalization of flow in the Paraíba do Sul River basin and its main tributaries. In this study, all available flow data were used to calculate the minimum, average and maximum characteristic flows (CPRM, 2003). After these determined flows, they were dimensioned to determine the homogeneous regions hydrologically and statistically.

With this methodology applied, the Pomba river basin was grouped with the Paraibuna Mineiro river basins (excluding the Preto basin), Pirapetinga and Angu, called the homogeneous VB region. The maximum annual flows were dimensioned by their respective average flood flow, Qmc (CPRM, 2003).

To determine the maximum flow rates in the recurrence times of 10, 25, 50 and 100 years, it is necessary to consult Table 12, below, multiplying the coefficient Q / Qmc by the flow estimated by the regional equation determined by CPRM.

(38)

38 Table 11: Dimensionless values of the maximum annual flows Q / Qmc, to its return period

RT (CPRM, 2003). RT (years) Q/Qmc Region VB 1,01 0,4197 1,10 0,6080 1,15 0,6548 1,20 0,6915 1,50 0,8312 1,80 0,9177 2 0,9623 3 1,1143 4 1,2112 5 1,2829 10 1,4936 15 1,6119 20 1,6945 25 1,7580 30 1,8096 35 1,8530 40 1,8905 45 1,9235 50 1,9529 55 1,9795 60 2,0037 65 2,0260 70 2,0466 75 2,0657 80 2,0836 90 2,1162 100 2,1453

(39)

39

For the calculation of maximum average flow (Qmc), Equation 13 (CPRM, 2003) is used:

𝑄_𝑀𝐶 = 0,2406𝐴0,8907(13)

Equation 13 has drainage areas between 142 km² and 8572 km² and an R² correlation coefficient of 0.9852. Table 13 shows the values calculated for maximum flows in the recurrence times of 10, 25, 50 and 100 years using the regional equation of CPRM.

Table 12: Maximum flow rates determined by the CPRM equation. CPRM

Code Station Drainage

Area (km²) Q/Qmc Qrt=10 years Qrt=25 years Qrt=50 years Qrt=100 years 58710000 Usina Itueré 784 91,0 136,0 160,1 177,8 195,3 58720000 Tabuleiro 322 41,2 _61,6 _72,5 _80,5 _88,4 58730001 Guarani 1650 176,6 263,8 310,5 345,0 379,0 58735000 Astolfo Dutra 2350 242,1 361,5 425,5 472,7 519,3 58750000 Piau 490 59,9 89,5 105,3 117,0 128,5 58755000 Rio Novo 835 96,3 _143,8 _169,3 _188,1 _206,6 58765001 Usina Maurício 1770 188,0 280,9 330,6 367,2 403,4 58770000 Cataguases 5880 547,9 818,3 963,2 1069,9 1175,4 58790000 Santo Antônio de Pádua 8210 737,6 1101,6 1296,6 1440,4 1582,3

Table 14 shows the percentage difference between the use of the CPRM regional equation and the regional equation developed in this thesis.

(40)

40 Table 13: Comparison of results between regional equations of CPRM and this thesis.

Deviation between this study and CPRM

Area (km²) Qrt=10 years Qrt=25 years Qrt=50 years Qrt=100 years 58710000 Usina Itueré 784 23% 25% 26% 27% 58720000 Tabuleiro 322 43% 48% 51% 53% 58730001 Guarani 1650 7% 8% 8% 8% 58735000 Astolfo Dutra 2350 -3% -3% -2% -2% 58750000 Piau 490 -2% 0% 1% 2% 58755000 Rio Novo 835 1% 3% 3% 4% 58765001 Usina Maurício 1770 19% 20% 22% 22% 58770000 Cataguases 5880 2% 2% 3% 3%

58790000 Santo Antônio de Pádua 8210 23% 24% 25% 25%

Analyzing the deviations between the regional equations, it is possible to verify that the flows estimated by the two equations were similar in 5 stations: Guarani, Astolfo Dutra, Piau, Rio Novo and Cataguases, with a percentage difference of up to 10%.

The stations Usina Itueré, Tabuleiro, Usina Maurício and Santo Antônio de Pádua showed marked percentage differences, mainly at the Tabuleiro station. These differences may be due to differences in methodology applied by the CPRM and by this thesis, input data errors or changes in the hydrological behavior upstream of the fluviometric stations.

In order to evaluate the latter hypothesis, the history of maximum flows of the last 17 years of these stations was analyzed and compared with the average of previous years, in order to assess the difference between them.

Table 14: Average maximum flow rates up to 1999 and the current century.

Area (km²) Average until 1999 (m³/s) Average after 1999 (m³/s) Deviation 58710000 Usina Itueré 784 103 144 -40% 58720000 Tabuleiro 322 57 94 -67% 58765001 Usina Maurício 1770 223 216 3%

(41)

41

Analyzing Table 15, it can be seen that the average maximum flow rates increased significantly in this century at the Usina Itueré and Tabuleiro fluviometric stations. Such a change in the flood record may be associated with several factors, such as greater susceptibility of the drainage area to climate change, change in land use and occupation, increased heavy rainfall, among others. It is recommended to monitor and verify the maximum flows in the next to confirm this tendency of increase and to carry out a specific study to determine its cause.

At the Usina Maurício station there was no significant change in the flows of the periods analyzed, whereas for the Santo Antônio de Pádua station it is not possible to carry out an analysis in the period after 1999, as there are only flows registered in 2 years. In these stations it was not possible to identify a probable cause in the different estimates of maximum flows through the regional equations of CPRM and this thesis, and the methodology applied in both may have generated significantly different values.

4.2 Flooding and risks

Floods can be generated by several factors, such as heavy rains, rupture of dams, tsunamis, among others, the first being the most common. Flow rates at different recurrence times present different risks for the flooded area.

Risk is the probability of an event occurring, whilst hazard is the severity of its consequences. The more likely the event to occur and the worse its consequences, the greater the risk.

Flow rates with shorter recurrence times (up to 10 years) have mild to moderate impacts on the micro drainage of an area, generating reversible impacts. Flow rates with recurrence times over 25 years can generate major floods and relevant economic impacts, especially if there are structural problems of macro-drainage and occupations of marginal areas of rivers.

Even more extreme flows, in times of recurrence above 100 years, can cause major disasters and lead to the collapse of large structures, such as bridges, dams and buildings, which can lead to a high number of losses of human lives.

Clarke (2002) draws attention to the need to take into account the occurrences of climate change and land use and occupation in the analysis of extreme events.

(42)

42

Milly el al. (2008) shows that anthropic actions are increasing the risk of flooding, actions such as river rectification, poorly dimensioned drainage works and changes in land use and coverage make areas more susceptible to flooding.

The European Environment Agency (2015) carried out a study that bought the increase in the registration of moderate, high and very high floods in the continent between 1980 and 2010 (Figure 3), resulting in a greater demand for resources for mitigation and prevention of these phenomena.

Figure 12: Reported flood phenomena between 1980 and 2010 (European Environment Agency, 2015).

Svetlana et al. (2015) points out the Germany, Italy and United Kingdom registering the highest economic losses in Europe,14,26 billion EUR,13,1 billion EUR, 6,4 billion EUR, respectively.

It is essential to elaborate more studies for flood modeling to be developed so that mitigation and prevention plans prevent losses of human lives and reduce their economic and social impacts.

(43)

43

5. Conclusion

The regionalization of maximum flow rates performed in this thesis proved to be an important tool, since with the regional equations generated can be used to estimate flow rates in the recurrence times of 10, 25, 50 and 100 years in places with little or no flow data. This information can be used by public bodies in the basin for flood studies, determining the most vulnerable areas.

In Brazil, flows of 10 and 25 years of recurrence time are used as flows of projects in micro-drainage works, as in dimensioning manholes and storm galleries, for example. Flow rates of 50 and 100 years of recurrence time are used as flow rates for projects in macro-drainage works, such as in the design of small hydroelectric power plant reservoirs, damping basins, containment dikes, among others.

There are several methodologies for regionalization of flows available, being necessary to evaluate which is the best according to the available data and the characteristics of the basin. The one chosen in this study was considered adequate, since the lowest coefficient of determination R² was 0.9368 and using only the drainage area as a variable, which data can be easily obtained.

In comparison with the results of other studies in other hydrographic basins, this study determined R² values that are comparable to those found in the literature, increasing the reliability of the generated regional equations.

Structural and non-structural measures proved to be important for flood prevention, in addition to planning policies, such as the project Hydrological Alert System of the Pomba River Basin. This project has good results, but there are still points to be improved, such as the elaboration of flood maps according to the recurrence time of flows.

The Pomba River has a drainage area of 8,616 km², but it was only possible to use nine fluviometric stations within its basin, explaining the chronic problem of the scant hydrological monitoring networks in the Brazilian basins. Greater investment in this sector is recommended to increase the hydrological knowledge of the Brazilian hydrographic basins, reducing the dependence on carrying out large studies of regionalization of flows.

(44)

44

After comparing the results generated with the flow estimates using the CPRM equation, it was found that the Usina Itueré, Tabuleiro, Usina Maurício and Santo Antônio de Pádua fluviometric stations showed significant differences.

Analyzes made on the average of the flows up to 1999 and in the current century showed a significant increase in Usina Itueré and Tabuleiro, possibly indicating a change in the hydrological behavior upstream of these stations.

In Usina Maurício and Santo Antônio de Pádua it was not possible to analyze or verify a change in the water regime, and the difference in methodology applied in this study and that of CPRM may have influenced the results, requiring future research to determine its cause.

In addition, there is a notable lack of maintenance and deactivation of stations over the years, as of the 102 initially listed, only 24 had some type of record. Therefore, better management and maintenance of the existing hydrological network is advised.

(45)

45

Reference list

Albuquerque Y., et al. (2020). Regionalization of minimum, average and maximum

flows in the Itapicuru River basin - BA. The Research, Society and Development journal.

Barth R. (1997). Master Plans in Urban Drainage: Proposal of Measures for its

Implementation. Planos Diretores em Drenagem Urbana: Proposição de Medidas para a sua Implementação. São Paulo University: Dissertation, 116p.

Bobée B.; Ashkar F. (1991). The Gama Distribution and Derived Distributions

Applied in Hydrology. Littleton, Colorado: Water Resources Press.

Bobée B.; Robitaille R. (1977). The Use of The Pearson Type 3 Distribution an Log

Pearson Type 3 Distribution Revisited. Water Resources Research.

Brazilian Federal Government (1997). Law n. 9.433, from January 8. Institutes the

national water resource policy.

Brazilian Water Agency (2010). Agência Nacional de Águas. Hidroweb System.

Available at: http://www.snirh.gov.br/hidroweb/apresentacao [Accessed on 10 Sep., 2020].

Brazilian Water Agency (2010). Agência Nacional de Águas. Study manual for

water availability and hydroelectric opportunities. Manual de estudos de disponibilidade hídrica para aproveitamentos hidroelétricos. Brasília, 73p.

Carmo S. (2015). Implementation of Ciliary Materials for the Recovery and

Rehabilitation of Water Resources. Implantação de Matas Ciliares para Recuperação e Reabilitação de Recursos Hídricos. Available at: https://oswaldocruz.br/revista_academica/content/pdf/Edicao_09_CARMO_S_Elaine _-_BONETTO_Nelson_Cesar_Fernando.pdf [Accessed on 20 Dec., 2020].

Castellarin A. (2007). Probabilistic envelope curves for design flood estimation at

ungauged sites. Water Resources Research, v. 43, n. 4.

Chow V. (1951). A General Formula for Hydrology Frequency Analysis. Trans. Am.

Geophysics. Um.

CLARKE R. (2002). Stochastic Hydrology Revisited. In: Revista Brasileira de

Recursos Hídricos, Vol. 7, No 4.

CLIMATE SERVICE CENTER (2013). The European Floods Directive and

Opportunities offered by Land Use Planning. Available at: https://www.climate-service-center.de/imperia/md/content/csc/csc-report_12.pdf [Accessed on 15 Nov., 2020].

Comitê para Integração da Bacia Hidrográfica do Rio Paraíba do Sul (2006).

(46)

46

da Bacia do Rio Paraíba do Sul – Resumo – Action Book Rio Pomba Basin. Rio de Janeiro, 113 p.

Coronado-Hernández O., et al. (2020). Selection of Hydrological Probability

Distributions for Extreme Rainfall Events in the Regions of Colombia.Multidisciplinary Digital Publishing Institute.

CPRM – Compania de Pesquisa de Recursos Minerais (2020). Hydrological Alert

System of the Pomba River Basin. Sistema de Alerta Hidrológico da Bacia do rio Pomba. Rio de Janeiro, 20 p.

CPRM – Compania de Pesquisa de Recursos Minerais (2017). Study of the 95%

streamflow permanence in the sub-basin 58. Estudo da vazão de 95% de permanência da sub-bacia 58. Rio de Janeiro, 136 p.

CPRM – Compania de Pesquisa de Recursos Minerais (2003). Summary report of

the work of Sub-basin Flow Regionalization 58. Relatório-síntese do trabalho de Regionalização de Vazões da Sub-bacia 58. Rio de Janeiro, 73 p.

CRUE (2008). Integrate, Consolidate and Disseminate European Flood Risk

Management Research: Systematization, evaluation and context conditions of structural and nonstructural measures for flood risk reduction. London, 160 p.

Eletrobrás (2020). Basic studies. Estudos Básicos. Available at: https://eletrobras.com/pt/Paginas/Manuais-e-Diretrizes-para-Estudos-e-Projetos.aspx [Accessed on 27 Sep., 2020].

EMBRAPA – Empresa Brasileira de Pesquisa Agropecuária (2011).Soil Map of Brazil. Mapa de Solos do Brasil. Rio de Janeiro, 67 p.

Euclydes H., et al. (2001). Hydrological Regionalization in the Upper São Francisco

Basin upstream of the Três Marias Dam, Minas Gerais. Regionalização Hidrológica na Bacia do Alto São Francisco a Montante da Barragem de Três Marias, Minas Gerais. Brazilian Journal of Water Resources. Available at: https://abrh.s3.sa-east-1.amazonaws.com/Sumarios/41/b434f785cf5722e309ccf07d37ff83a8_875a16e3192 61b113c7c71f4b232878c.pdf [Accessed on 20 Jan., 2021].

European Commission (2003). Best Practices on Flood Prevention, Protection and

Mitigation. Available at:

https://ec.europa.eu/environment/water/flood_risk/pdf/flooding_bestpractice.pdf [Accessed on 10 Nov., 2020].

European Environment Agency (2015). European Past Floods. Available at:

http://www.eea.europa.eu/data-and-maps/data/european-past-floods [Accessed on 20 Fev., 2021].

Geo Water Resource Brazil (2007). Geo Brasil Recursos Hídricos. Component of

the series of reports on the State and perspectives of the environment in Brazil. Componente da série de relatórios sobre o Estado e perspectivas do meio ambiente no Brasil. Brasília, 264 p.

(47)

47

Guedes S., et al. (2009). Study of the self-purification capacity of the Rio Pomba

using the QUAL2KW model. Estudo da capacidade de autodepuração do Rio Pomba utilizando o modelo QUAL2KW. Campo Grande: Simpósio Brasileiro de Recursos

Hídricos. Available at:

https://www.abrh.org.br/SGCv3/index.php?PUB=3&ID=110&PUBLICACAO=SIMPOS IOS[Accessed on 20 Jun., 2020].

Haberlandt U., et al. (2006). Regionalization of the base flow index from dynamically

simulated flow components — a case study in the Elbe River Basin. Journal of Hydrology.

IBGE – Brazilian Institute of Geography and Statistics (2010). Instituto Brasileiro

de Geografia e Estatística. Brazilian Demographic Census 2010. Censo Demográfico 2010. Available at: https://censo2010.ibge.gov.br/ [Accessed on 10 Sep., 2020].

IPCC - Intergovernmental Panel on Climate Change (2012). Special Report on

Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation.

ITAIPU BINACIONAL (2010). Dam Type. Tipo da Barragem. Available at:

https://www.itaipu.gov.br/ [Accessed on 17 Nov., 2020].

Kapoor R. (2020). Sea Wall in The Maldives and Its Sustainability. Available at:

https://maritimeindia.org/sea-wall-in-the-maldives-and-its-sustainability/ [Accessed on 20 Nov., 2020].

Köeppen W. (1948). Climatologia: Con un estudio de los climas de la tierra. México

Buenos Aires: Fondo de Cultura Económica.

Laaha G.; Blöschl G. (2006). A comparison of low flow regionalization methods -

catchment grouping. Journal of Hydrology.

Lee H., et al. (2005). Selection of conceptual models for regionalization of the

rainfall-runoff relationship. Journal of Hydrology.

Lopes T., et al. (2017). Regionalization of Maximum and Minimum Flow in the Teles

Pires Basin, Brasil. Available at:

https://www.scielo.br/scielo.php?script=sci_arttext&pid=S0100-69162017000100054 [Accessed on 20 Jan., 2021].

Mamum A., et al. (2006). Regionalization of low flow frequency curves for the

Peninsular Malaysia. Journal of Hydrology.

Matos T., et al. (2020). Regionalization of maximum, minimum and mean

streamflows for the Juruena River basin, Brazil. Available at: https://www.scielo.br/scielo.php?pid=S1980-993X2020000300300&script=sci_arttext [Accessed on 20 Jan., 2021].