One Ecosystem :
Research Article
|
Corresponding author: Eva Horváthová (ev.horvathova@gmail.com)
Academic editor: Alessandra La Notte
Received: 11 Feb 2022 | Accepted: 16 Sep 2022 | Published: 10 Nov 2022
© 2022 Eva Horváthová
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Horváthová E (2022) Analysis of Drinking Water treatment costs – with an Application to Groundwater Purification Valuation. One Ecosystem 7: e82125. https://doi.org/10.3897/oneeco.7.e82125
|
|
Understanding the factors affecting drinking water production costs is crucial for choosing a cost-effective solution for public drinking water supply systems. An important determinant of water treatment costs is the purification of raw water. Despite water purification being a well-acknowledged ecosystem service, its monetary value has not been assessed much yet. We present the first study analysing the determinants of drinking water production costs and valuating groundwater purification in the Czech Republic. We tested the impact of the type of raw water, the amount of drinking water produced, electric power consumption and treatment technologies and chemicals. The results suggested that drinking water production from groundwater was cheaper than from surface water. Even though drinking water production from groundwater was cheaper than from surface water, the application of some technologies, for example, chlorine or manganese removal, increased the production cost. Hence groundwater production costs can exceed surface water production costs. The outcome of the regression was applied for the valuation of groundwater purification. The valuation was further used for the development of monetary drinking water accounts within the System of Environmental- Economic Accounting – Ecosystem Accounting.
drinking water, groundwater, replacement cost method, treatment costs, System of Environmental-Economic Accounting – Ecosystem Accounting (SEEA EA)
Water treatment cost depends on raw water quality, treatment technologies, regulations, energy source and the amount of water treated (
Water purification is probably amongst one of the most cited benefits provided by nature to humans, so-called ecosystem services (
Despite there is a growing demand for water resources valuation (
The replacement cost method estimates the economic value of an ecosystem service by the cost of replacing the service with a man-made substitute (
Furthermore, the previous research on determinants of drinking water production costs focused mainly on North America and western Europe. In central Europe, there is a lack of water valuation studies, with a notable exception for the valuation of green water (
Therefore, we analysed drinking water production costs in a central European country - the Czech Republic. Next, we applied the results for the valuation of groundwater purification and development of the SEEA EA monetary drinking water accounts. To the best of our knowledge, this is the first study to investigate the determinants of drinking water production costs and to value groundwater in central Europe. Further, we believe that the SEEA EA monetary drinking water accounts have not been developed in central Europe yet.
A database was obtained by merging data that the owners and operators of water supply systems mandatorily submit to the Ministry of Agriculture. The database contains one-year property and operating data (2018) for all water withdrawal points in the Czech Republic from which drinking water is commercially produced. To monitor compliance with pricing rules, the owners and operators of water supply and sewerage systems are obliged to submit a comparison of all items of the price calculation to the Ministry of Agriculture every year. The items are defined in the implementing decree to the Water Supply and Sewerage Act (
Drinking water is produced either from groundwater or from surface water in the Czech Republic. Groundwater is an important source of drinking water amounting to about 49% of the total drinking water production with a constant share during the last years (
A generic short-run cost function of a firm using an environmental input (
C = f (WP, X, N, F, E) (1)
where WP is the amount of output, X are firm-specific characteristics, for example, data on management, N is costs of non-environmental inputs, for example, labour, energy and treatment chemicals, F is costs of fixed factors and E is the quality of the natural capital.
As we investigated a unit change, we used a linear cost function instead of a logarithmic one which is used when elasticities are studied. Next, as we were particularly interested in the impact of the type of raw water on the unit production costs, we estimated the following linear cost function:
UCWC = β0+ β1lnWP + β2Power + β3Groundwater_d + β4NoSludgeTreat + β5 NoTreatment + b1 TreatmentTechnology1 + b2 TreatmentTechnology2 +……..+ b30 TreatmentTechnology30 + e (2)
where UCWC is a unit cost without charges paid for water withdrawal; WP is the amount of water produced; Power is a unit consumption of electric power; Groundwater_d, NoSludgeTreat and NoTreatment is a dummy for a type of raw water, sludge treatment and no treatment, respectively. TreatmentTechnology are dummies for treatment technologies and chemicals. To account for non-linearities in output level as was shown in previous studies (
Variable | Description | Obs | Mean | Std. Dev. | Min | Max |
UCWC | Unit cost without charges (CZK/m3) | 3,253 | 12.73 | 9.77 | 0.52 | 49.9 |
WP | Total amount of water produced (km3/year) | 3,253 | 176.1 | 1,820 | 0.02 | 87,157 |
Power | Unit consumption of electric power (kWh/m3 water produced) | 3,253 | 0.71 | 1.36 | 0 | 43.64 |
Groundwater_d | Dummy v., = 1 if groundwater/total amount of water produced >= 0.5 | 3,253 | 0.96 | 0.2 | 0 | 1 |
NoSludgeTreat | Dummy v., = 1 if no sludge treatment | 3,253 | 0.36 | 0.48 | 0 | 1 |
NoTreatment | Dummy v., = 1 if no water treatment is applied | 3,253 | 0.55 | 0.5 | 0 | 1 |
Deacidification | Dummy v., = 1 if Deacidification by filtration or aeration applied | 3,253 | 0.1 | 0.3 | 0 | 1 |
Demanganisation | Dummy v., = 1 if Demanganisation applied | 3,253 | 0.105 | 0.306 | 0 | 1 |
Filtration | Dummy v., = 1 if Filtration applied | 3,253 | 0.165 | 0.372 | 0 | 1 |
ChemDisinfection | Dummy v., = 1 if Chemical disinfection applied | 3,253 | 0.378 | 0.485 | 0 | 1 |
Chlorine | Dummy v., = 1 if Chlorine applied | 3,253 | 0.106 | 0.308 | 0 | 1 |
IronRemoval | Dummy v., = 1 if Iron removal applied | 3,253 | 0.117 | 0.322 | 0 | 1 |
OtherAggregation | Dummy v., = 1 if Other aggregating agent applied | 3,253 | 0.075 | 0.263 | 0 | 1 |
OtherTechnology | Dummy v., = 1 if Other technology applied | 3,253 | 0.074 | 0.262 | 0 | 1 |
PotassiumPermangan | Dummy v., = 1 if Potassium permanganate applied | 3,253 | 0.063 | 0.244 | 0 | 1 |
RadonRemoval | Dummy v., = 1 if Radon removal applied | 3,253 | 0.079 | 0.27 | 0 | 1 |
SodiumHypochlorite | Dummy v., = 1 if Sodium hypochlorite applied | 3,253 | 0.872 | 0.334 | 0 | 1 |
We had no data for firm-specific characteristics (X), as well as for costs of fixed factors (F). Costs of non-environmental inputs (N) were represented by the consumption of electric power (Power) and sludge treatment (NoSludgeTreat). Quality of the natural capital (E) was represented by the type of raw water (Groundwater_d), no treatment (NoTreatment) and dummies for treatment technologies and chemicals (TreatmentTechnology).
First, we calculated unit costs without charges. Since the water production cost included the charges paid for raw water and the charge rate was locally and type specific, we deducted them from the water production cost. We calculated the unit costs without charges for an abstraction point a (UCWCa) as follows:
UCWCa=(TPCa-SWa*CRSa-GWa*CRG)/WPa (3)
where TPCa is total production costs for an abstraction point a, SWa is the amount of surface water abstracted at abstraction point a, GWa is the amount of groundwater abstracted at abstraction point a, CRSa is charge rate for surface water applied at abstraction point a, CRG is charge rate for groundwater and WPa is the amount of drinking water produced at abstraction point a.
Since the TPCa were not included in the database, we calculated them as:
TPCa=UCa*WPa (4)
where UCa is the unit production costs (CZK/m3). UCa was included in the database for each abstraction point. The drinking water producers calculate UCa as:
UCa=TPCa/IWa (5)
where TPCa is the sum of material costs (raw water, chemicals, other material costs), energy costs, wages and salaries, other direct costs (depreciation; repair, rent and renovation of infrastructure assets), operating costs, financial costs and overhead costs. IWa is the amount of invoiced drinking water. We assumed that the amount of water produced equals the amount of invoiced water. The difference in the amount of water produced and invoiced can be either caused by storage or leakages. The companies store water to balance differences between the demand for water and its production. Hence, the stored water is distributed and invoiced in the next year similarly as stored water produced in the previous year was distributed in the studied year. Costs of leakages and the water leakage prevention costs are included in the total production costs.
Next, we cleaned the data in the database. We dropped observations with too low water production (the total amount of water produced < 0.01 km3/year) and too low (1 m3 < 0.5 CZK) or too high (1 m3 > 50 CZK) UCWC (the thresholds for dropping observations were discussed at the Ministry of Agriculture - data provider). We supposed that too high or too low costs were entered wrongly. We also dropped three abstraction points where more than 50% of water production accounted for technological water. Next, we dropped five observations where infiltration was reported because infiltration is applied on one site only in the Czech Republic. Hence, the four observations were entered wrongly. After the data cleaning, 3,253 observations remained (the total number of observations before the changes was 3,566).
First, the value of the groundwater purification (GPV) was calculated as:
GPV= WPG * ß3/ 26.444 (6)
where WPG is the amount of drinking water produced from groundwater (stated in the database) and ß3 (calculated in the egation 2) is the difference in production costs of drinking water from groundwater and surface water. The coefficient for Groundwater_d measures the average difference in production costs between groundwater and surface water when other factors (i.e. the amount of water produced, unit consumption of electric power, usage of treatment technologies and chemicals, sludge and no treatment), have the same levels (
Next, accounting tables according to the SEEA-EA were set up. Use tables record a flow of an ecosystem service to beneficiaries while supply tables depict which ecosystem types supply the ecosystem service (
First, we estimated a full model including all the explanatory variables, i.e. the ln WP, Power, Groundwater_d, NoSludgeTreat, NoTreatment and dummies on the 13 treatment technologies and chemicals. Since heteroscedasticity was detected (Breusch-Pagan test: F(16, 3236) = 5.35, Prob > F = 0.0000), robust errors were calculated for all specifications. The coefficients were statistically significant for the logarithm of the amount of water produced, electric power consumption and dummies for groundwater and some treatment technologies and chemicals (demanganisation, chemical disinfection, chlorine, other aggregating agent and sodium hypochlorite).
Next, we dropped a variable with the lowest absolute value of the t-statistic (the following variables were gradually dropped: NoTreatment, NoSludgeTreat, PotassiumPermanganate, Filtration, RadonRemoval, Deacidification, OtherTechnology and IronRemoval) to simplify the model until significant variables only remained. In total, nine model specifications were tested and the results of all these specifications are reported in Suppl. material
Foremost, we found that the companies which produce drinking water mainly from groundwater experienced significantly lower production costs compared to the companies which produce drinking water mainly from surface water. The magnitude of this effect depended on the specification of the model and it ranged between 2.08 and 2.47. Hence, the drinking water production unit costs were 0.078 - 0.093 EUR lower if the drinking water was produced from groundwater.
Then, we confirmed economies of scale as the unit water production cost without charges significantly decreased with the logarithm of total water produced. This finding was significant at a level of 1% in all tested specifications. Next, the unit production costs slightly increased with the unit consumption of electric power (0.018-0.019 EUR/m3).
Lastly, some treatment technologies and the application of some chemicals (demanganisation, chemical disinfection, chlorine, other aggregating agent and sodium hypochlorite) increased production costs. The highest impact occurred when sodium hypochlorite, chlorine and demanganisation were applied, which increased production unit costs by 0.179 – 0.181 EUR, 0.181-0.188 EUR and 0.102 – 0.150 EUR, respectively.
The difference in production costs of drinking water from groundwater and surface water depended on the model specification. It ranged between 0.078 and 0.093 EUR/m3. For the groundwater purification valuation, we used the cost difference of the model with significant variables only, which was 0.085 EUR/m3 (the unit value of the service). The amount of drinking water produced from groundwater was 274,032 km3 in 2018. Hence, the value of groundwater purification was 23.16 M EUR.
Next, monetary supply and use tables were set up (Tables
Ecosystem types |
||||||||||
Measurement units |
Cropland |
Woodland and forest |
Urban |
Heathland and shrub |
Grassland |
Rivers and lakes |
Wetlands |
Sparsely vegetated areas |
TOTAL SUPPLY |
|
Ecosystem service |
||||||||||
Groundwater purification |
K EUR/year |
9,533 |
6,746 |
2,626 |
2,274 |
1,806 |
274 |
31 |
2 |
23,293 |
Total |
23,293 |
Institutional sector |
||||||||||||
Measurement units |
agriculture |
forestry |
fisheries |
mining and quarrying |
manufacturing |
construction |
electricity, gas supply |
water collection, treatment, supply |
other industries |
households |
TOTAL USE |
|
Ecosystem service |
||||||||||||
Groundwater purification |
K EUR/year |
23,293 |
23,293 |
|||||||||
Total |
23,293 |
23,293 |
The Supply Table (Table
It is necessary to understand the factors affecting drinking water treatment costs for designing cost-efficient public water systems. Moreover, the monetary valuation of the groundwater purification services, which have not been assessed to a great extent yet, would help to improve decision-making processes. This paper contributes to the existing literature by analysing the determinants of the drinking water production costs and estimating the value of the groundwater purification service in the Czech Republic.
The results showed that drinking water production from groundwater was cheaper than from surface water. However, some treatment technologies increased the treatment costs; hence, drinking water production from groundwater can be more expensive than from surface water if these technologies have to be applied. Next, we confirmed the economies of scale in drinking water production, which implies that centralised water treatment is more cost-efficient. Decreasing drinking water production costs with the logarithm of the amount of water treated were shown in previous studies (
The estimated cost function was similar to a generic cost function, but we lack data for some explanatory variables, for example, data on firms´ characteristics and fixed factors. Next, we also had limited data on the costs of non-environmental inputs. As site-specific factors have the highest impact on the drinking water production costs (
The R-squared value for all specifications was relatively low (0.08). However, there is no assumption about a minimum level of R2 in linear regression models. Low R2 just means that a low amount of variation in the dependent variable is explained by the independent variables (
Overall, the results suggested that drinking water production from groundwater was cheaper than from surface water. This is due to a usually better quality of groundwater relative to surface water (
To assess the monetary value of the groundwater purification, we used regression results for the valuation of the groundwater purification service by the replacement cost method. The replacement cost method was applied for the valuation of the purification of surface water (
The smaller cost difference in our study (0.078-0.093 EUR/m3) was probably on account of controlling for other variables. The Dutch studies compared the average production costs only and failed to control for key variables, such as treatment technologies, electric power consumption and economies of scale. As a part of the cost difference can be probably attributed to these variables, the higher cost differences in the Dutch studies were probably caused by the omitted variables.
It should be emphasised that this approach measures the value of an extra-purification of groundwater relative to surface water only. The value of purification of surface water is not calculated even though its value is substantial (
The valuation results were used for the development of monetary groundwater purification supply and use tables within the SEEA EA framework. A 5 km-wide buffer zone around each groundwater withdrawal point was used for the supply table compilation as no detailed models of groundwater flows were available. The buffer zone approach is often used in the absence of groundwater flow models (
Continued research is needed to improve estimated relationships. First, a panel data analysis would help to mitigate the problem of missing companies´ characteristics. Likewise, more research is needed to quantify relationships between treatment costs and landscape characteristics as the links between ecosystem types and water quality are well established (
We thank Věra Bogdanova (Ministry of Agriculture of the Czech Republic) and Ondřej Lípa (Ministry of Agriculture of the Czech Republic) for providing data and consultations.
This research has been supported by the Czech Academy of Sciences, programme Strategy AV21 (project No. 21 “Záchrana a obnova krajiny”) and by project MAIA (Mapping and Assessment for Integrated ecosystem Accounting), EU call H2020-SC5-2018-1, Grant Agreement No. 817527.
Eva Horváthová: Conceptualisation, Data curation, Formal analysis, Methodology, Writing
The author declares that she has no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Regression results. Dependent variable: UCWC (the unit costs without charges)
Ecosystem types (%) in groundwater buffer zones (Supply account) and the whole Czech Republic.
Excluded technologies include AC filtration, Activated powdered carbon, Al-based destabilising agent, gating agent, Biological filtration, Clarification, CO2, Coagulation, Denitrification, Fe-based stabilising agent, Chlorine dioxide, Ion exchange, Lime hydrate, Membrane filtration, Ozone, Ozonisation, Sedimentation, Sodium carbon, Sodium hydroxide, Stabilisation and UV radiation.