Leachate: production and characterizaton

G. J. FARQUHAR
Department of Civil Engineering, University of Waterloo, Waterloo,
Ont., Canada N2L 3G1
Received February 8, 1988
Revised manuscript accepted December 2,1988

Leachate production deals with the creation of contaminated liquid at the base of a landfill. It involves the elements of a water balance in which precipitation either runs off from the landfill or infiltrates. Some infiltration will evapotranspire, some may be stored within the landfill, and the balance becomes percolate and eventually leachate. Examples of the water balance method are presented and consideration is given to the sequence of landfill development and the effect that it has on leachate flow.

Information is provided on the microbial decomposition of municipal solid waste and the conditions which influence decomposition rates. The impact of the microbial processes on leachate contaminant composition is examined. Leachate characterization requires that estimates be made of contaminant types and concentrations as a function of refuse age. Data describing contaminant production from various sources are examined. A procedure is presented to combine the water balance method and contaminant production curves to predict leachate flow and strength with respect to site age. A simpler alternative involving a set of tables showing expected contaminant types and ranges of concentrations as a function of refuse age is also provided.

Key words: landfill, leachate, flow, composition, estimation, model.

La production d'eaux de ruissellement entraine la formation de liquides contaminés á la base des sites d'enfouissement. Ce phénoméne met en cause les éléments du bilan hydrologique qui font en sorte que les précipitations s'écoulent du site d'enfouissement ou s'y infiltrent. Une certaine quantité des eaux d'infiltration s'élimine par évapotranspiration, alors qu'une autre partie s'accumule dans le sol du site d'enfouissement; le reste s'infiltre par les interstices du sol pour atteindre la nappe phréatique. Des exemples de la méthode du bilan hydrologique sont présentés et une attention particuliére est accordée á la séquence de développement du site d'enfouissement et á son effet sur l'écoulement des eaux de ruissellement. Des renseignements sont fournis sur la décomposition microbienne des déchets solides et les conditions qui influent sur la vitesse de décomposition. L'impact du processus microbien sur la composition des matiéres contaminantes des eaux de ruissellement est examiné. La caractérisation de ces eaux de ruissellement exige que des évaluations des types et des concentrations de contaminants soient faites, en fonction de l'age des détritus. Des données décrivant la production de contaminants á partir de diverses sources sont examinées. Une procédure permettant de combiner la méthode du bilan hydrologique et les courbes de production des contaminants afin de prédire l'écoulement et la concentration des eaux de ruissellement en fonction de l'age du site est présentée. Un ensemble de tableaux montrant les types de contaminants prévus et les échelles de concentration en fonction de l'age des détritus est également fourni.

Mots clés: site d'enfouissement, eaux de ruissellement, écoulement, estimation, modéle.

Can. J. Civ. Eng. 16.317 - 325 (1989) [Traduit par la revue]

Introduction

Landfills have served for many decades as ultimate disposal sites for all manner of wastes: residential, commercial, and industrial, both innocuous and hazardous. Landfill technology has evolved from the open, burning dump to highly engineered sites designed to minimize tile impact of contaminants in the waste on the adjacent environment.
The major environmental problems experienced at landfills have resulted from the loss of leachate from the site and the subsequent contamination of surrounding land and water. Improvements in landfill engineering have been aimed primarily at reducing leachate production, collecting and treating leachate prior to discharge, and limiting leachate discharge to the assimilative capacity of the surrounding soil. Whether leachate is to be collected and treated or is allowed to discharge to the soil, it is essential to have estimates of leachate flow and strength and the variation of these with time as the site develops, through closure and after closure. While these estimates are essential to proper landfill design, their preparation is a difficult and uncertain process. This paper examines current methods and data available for estimating leachate production and variability.
________________________
Note: Written discussion of this paper is welcomed and will be received by the Editor until October 31, 1989 (address inside front cover).

Mechanisms of leachate formation

Leachate is produced when moisture enters the refuse in a landfill, extracts contaminants into the liquid phase, and produces a moisture content sufficiently high to initiate liquid flow. Sources of moisture entering the landfill include liquid present in the refuse at placement, precipitation falling on refuse at placement and infiltrating after cover application, and intrusion of groundwater from outside into the landfill.

A generalized pattern of leachate formation is presented in Fig. 1. The components shown include the following steps:

1. Precipitation (P) falls on the landfill and some of it becomes runoff (RO).

2. Some of P infiltrates (I) the surface (uncovered refuse, intermediate cover, or final cover).

3. Some of I evaporates (E) from the surface and (or) transpires (T) through the vegetative cover if it exists.

4. Some of I may make up a deficiency in soil moisture storage (S) (the difference between field capacity (FC) and the existing moisture content (MC)).

5. The remainder of I, after E, T, and S have been satisfied, moves downward forming percolate (PERC) arid eventually leachate (L) as it reaches the base of the landfill.

6. PERC may be augmented by infiltration of groundwater (G). The procedure used to analyze these processes is referred to as a water balance (WB), various forms of which are commonly

used for the simulation of surface water hydrology. The algebraic statement of this form of water balance is

[1] PERC = P - RO - ET - AS + G

While [1] is conceptually correct and comprehensive, accurate predictions of leachate flow are difficult to achieve because of the uncertainties associated with estimating the various terms. Most formulae and methods in use are empirical. Some of the data base required is stochastic in nature (temperature, heat index, precipitation, wind, vegetative growth). Other data are poorly defined (runoff coefficients, refuse arid cover density and compaction, moisture storage capacities).

Analyses have been performed to compare water balance predictions of leachate flow with actual measurements made in the field. Gee (1981) used two variations of the water balance method to predict leachate flow at the GROWS Landfill in Bucks Co., Pennsylvania. These predictions were too high by a factor of approximately 2 when compared with measured leachate flows. Lu et al. (1981) performed similar comparisons at 5 landfills using 25 different methods to estimate the various terms of [1]. On average, leachate flow estimates were in error by a factor of 2. However, the poorest estimates were as much as 100 times greater than the measured leachate flows.

In contrast, Kmet (1982) had excellent success using a water balance method to simulate leachate production in Ham's (1980) eight field lysimeters. Leachate flows ranged from 16.6 to 22.1% of precipitation on an annual average basis. Water balance methods predicted an average of 22% of precipitation, providing excellent agreement with measured values. Kmet used the water balance method (WBM) proposed by Fenn et al. (1975) with modifications to account for infiltration and runoff from the landfill during winter conditions. This appears to be an acceptable procedure to predict leachate flow.

The WBM as proposed by Fenn et al. (1975) is a manual procedure solved generally with monthly averaged values. Computer models have been developed subsequently using the WBM as a basis with various modifications. The hydrologic simulation of solid waste disposal sites (HSSWDS) model developed by Perrier and Gibson (1981) and the hydrologic evaluation of landfill performance (HELP) model reported by Schroeder (1983) are two of the more widely accepted of these computer models. The HELP model is perhaps the best of the available computer models. Its use has become compulsory for Superfund Site evaluation.

Components of the WBM and the HELP model
The component steps of the WBM used to calculate landfill leachate flow are presented in flow chart form in Fig. 2. Table 1 provides a summary description of each step and compares the steps with

those of the HELP model. More detailed information about the methods is given in Kmet (1982, 1986). Tile components are generally calculated in units of height of water per unit time (e.g., centimeters (or inches) per month).

The time of leachate arrival at the base of the landfill is handled differently in the two methods. The WBM does not account for the period of time required for the refuse to be brought up from its moisture content (MC) at placement (e.g., 15 cm.m-¹) to field capacity (FC) (e.g., 35 cm.m-¹) at which point liquid flow begins. This can take several months depending on the refuse type, compaction, and depth in addition to the percolation rate. The WBM assumes that the refuse is already at FC and that a unit of PERC at the top produces an equivalent unit of leachate at the bottom. The method is therefore applicable only after FC has been reached in the landfill. It is possible to estimate the time of first leachate production using PERC and allowing it to increase the MC until FC is reached. However, Kmet (1986) has shown that the actual time at which leachate first appears at a landfill is much less than that predicted with this method. He attributed this to channelling and nonhomogeneous MCs and flow properties within the refuse. This would result in FC in the region of channels much less than FC at other locations within the refuse.

The HELP model, in addition to performing a water balance, also calculates a flow rate through the refuse and therefore estimates the time of first leachate appearance. However, channelling due to heterogeneities within the refuse reduces the accuracy of these flow rate calculations and tends to overestimate leachate arrival times.

FARQUHAR

Table 1. Components of the WBM and comparison with the HELP model
============================================================
                                  WBM                                                             HELP model
      _____________________________________________________________________

        1. Pontential evapotranspiration (PET)-the potential amount of moisture that can evaporate
            soil and (or) refuse and transpire from vegetative cover depending upon temperature (T)
           and solar radiation
           Thornthwate method                                            Penman method (Penman 1948)
           (Thornthwaite and Mather 1957)
            Data input
              Temperature: Monthly                                       Daily
              Heat index: Monthly                                          Monthly
        2. Precipitation (P) - the precipitation in all forms falling on the site
            Monthly averages                                                Daily averages (choice available)
        3. Runoff (RO) - that portion of precipitations which runs off the site and does not infiltrate
            Thornthwaite and Mather (1957)                         U.S.Department of Agriculture (1975)
            Fenn et al. (1975)
            RO = C _ROP                                                        Same as WBM
            (C_RO is a runoff coefficient)
            Data input
               Vegetation
               Slope
               Soil types
               Surface treatment
       4. Infiltration (1) - that portion of precipitation which infiltrates the site
           I = P - RO                                                          Same as WBM
       5. Soil moisture storage (S) -the amount of infiltration which is retained in the soil and (or)
           refuse up to field capacity and thus does not percolate as leachata (SMAX is the maximum S
           for soil or refuse)
          Thornthwaite and Mather (1957)                          Same as WBM
          Data input
              Soil type
              Depth
              Field capacity
              Moisture content (MC)
              Wilting point (WP)
              AS (change in S)
                 = + ve when 1> 0 and MC < SMAX
                 = - ve when (1 - PET) <0 and
                                      MC> WP
         6. Actual evapotranspiration (ET) - the actual amount of evapotranspiration (ET < PET) which
             occurs; depends on the soil types and depths, vegetation type, root depth, MC
            ET < PET - [(I - PET) -   S]
        7. Percolation (PERC)- the amount of liquid which reaches the base of the landfill to become
            leachate PERC = P -RO - ET -   S + G

Output of the WBM
The output from the WBM and also the HELP model (although the Iatter is more comprehensive) is represented in Fig. 3. The data are given as monthly averages and thus some seasonality is evident. Leachate production (PERC) exists in the months from February to May inclusive and is zero during the remaining months. Seasonality in the other components such as P and ET is also shown. When PET exceeds I, S is reduced by evapotranspiration as shown in the months of June, July, and August. In November and December, although I exceeds PET no PERC occurs since l is used to make up the deficit in S (SMAX = 9 cm in this illustration).
The output therefore provides estimates of leachate flow and its variability throughout the year. Such information is essential when assessing the impact of the leachate on either the soil environment or a treatment facility to which the leachate is being discharged.

Predicting leachate flow under changing landfill conditions

An assumption inherent in the previous section is that the landfill conditions remain constant during the analysis. While this is essential to the analysis, it must be acknowledged that the landfill conditions are not uniform throughout the landfill at any time and that these conditions will change as the site ages. Some sections of the landfill may have final cover while others may have intermediate cover or no cover at all as at the working face. Both the area and the depth of the landfill will increase with time. As well, at any time, the refuse in the landfill ranges in age from new to old and will therefore have been exposed to different amounts of PERC. Field-scale calculations of leachate flow must take these variations into account.

Figure 4 shows a schematic representation of landfill development and leachate flow rate change as the landfill grows in size and depth with time. The leachate flow rate increases as the surface area of the landfill increases. The flow rate decreases after the application of less permeable soils and vegetation for the final cover.

The impact of leachate on the environment or on a system for collection, treatment and disposal will be influenced by these variations in leachate flow. Consequently, the WBM or HELP model simulations must be done sequentially to account for the changing conditions throughout the life of the landfill.

The leaching process

As PERC moves through the refuse, contaminants are mobilized into the liquid phase through dissolution and suspension from the stationary refuse phase, thus producing a contaminated leachate. Increased moisture enhances microbial activity within the landfill. As a result, metabolic by-products such as volatile fatty acids and alcohols are contributed to the leachate, increasing its organic strength. Some organic compounds augment the leaching potential of the liquid because of increased acidity and complexing potential.

Table 2 provides information on the composition of municipal solid waste based on several analyses performed throughout Canada and the United States. The component compositions are presented as ranges of percent wet weight of refuse. Category A consists of readily biodegradable food and garden wastes which produce high concentrations of organic matter (as BOD or TOC) and total Kjeldahl nitrogen in the leachate. This often occurs within the first few months of leaching. Category B is also organic but less biodegradable than A. It includes primarily newsprint and other paper with much smaller amounts of wood and rubber as examples. Because of reduced biodegradability, these components yield organics to the leachate at concentrations much lower than for Category A but for much longer times measured in years.

Fig. 4. Landfill progression

Table 2. Typical municipal solid waste composition*
                                          

                                                                                        Composition
                                                                                              range
                                          Componet                              (% wet weight)
                                        _____________________________________
                                        A. Food                                          5-20
                                             Garden residue                          15-25

                                        B. Paper                                         40-50
                                            Plastic                                           1-5
                                            Wood                                           2-5
                                            Other organics                              2-10

                                        C.Iron                                              5-10
                                            Other metals                                 0-1
                                      
                                        D.Glass                                             5-10
                                            Other inorganics                            2-5

                                            Moisture content                           20-30
                                       ___________________________________
                                       *Rovers and Farquhar 1973; Emcon Associates
                                        1975; McGinley and Kmet 1984.

Category C includes metallic wastes composed mainly of iron, aluminum, and zinc. In time, these and other metals appear in the leachate and do so for many years because of slow rates of release.

Category D includes nonmetallic inorganic components such as glass, soil, and salts. The readily soluble of these appear in the leachate in the first few months of leaching, while the less soluble will yield contaminants for several years. The alkaline earth metals (calcium, magnesium, sodium, and potassium) and the common anions (chloride, sulphate, phosphate, and carbonate) arise mainly from these waste components. In all cases, there is a limit to the amount of contaminant that can be leached from the refuse.

Because of these trends and conditions, the concentration of most contaminants in the leachate varies with time. Most contaminants, especially biodegradable organics, tend to reach peak concentrations in the leachate in the earlier months of leaching and then reduce subsequently. However, some contaminants such as poorly biodegradable organics and iron tend to persist in the leachate for several years. This is shown in a generalized way in Fig. 5. Information about typical leachate contaminant types and concentration ranges as they change with time is presented subsequently.
Figure 4 shows a schematic representation of landfill development with time in which each year's refuse deposition is identified by a number. At any time, for example during the 10th year, some refuse is fresh and not leached at all, while other refuse has been in place and exposed to leaching for 10 years. Each year's refuse will have a different age and thus will be at a different point on the time axis in Fig. 5. The older refuse will be producing leachate with contaminant concentrations represented by the right-hand side of Fig. 5, while the left-hand side applies to younger leachates. Thus the leachate produced in the 10th year will have contaminant concentrations which are weighted averages. They will be averaged from different sections of the landfill having refuse of different ages and different leaching histories. It is also apparent from this analysis that contaminants are contributed to the leachate for many years after the site is closed.

Microbial degradation of refuse in landfills

Except for the first few days after refuse placement, microbial decomposition in landfills proceeds under anoxic conditions. Hydrolytic and fermentative microbial processes solubilize waste components, producing organic acids, alcohols, ammonia, and carbon dioxide as major products. These processes are vigorous and rapidly initiated as the moisture content increases in the landfill. After several months, methanogenesis is initiated with methane and carbon dioxide produced as by-products. Methanogenesis is a slower and more fastidious process.
Microbial degradation in landfills has been described in detail by Farquhar and Rovers (1973), Fungaroli and Steiner (1979), Zehnder (1978), and Ham (1980) and will not be examined in detail here. Two concepts are important to establish, however:

Table 3. Conditions affecting methanogenesis in landfills

                                                               Optimal                            
                                                                  value                   Representative ranges for
                           Parameter                 (Zehnder 1978)         southern Ontario landfills*
                        _______________________________________________________
                        Temperature (°C)                  35                     10-20 at depth > 2 m
                pH                                          7.2                   5.5-6.5 for young leachate
                Moisture content            Saturation                 20-30 at placement
                          (%wet weight)
                       Oxidation reduction             < - 330                 > -330 at placement
                 potential (mV)
                       Nutrients                          Sufficient                   Phosphorus deficient
                       Toxicants                         None                        Ammonia and certain metals
                       _________________________________________________________
                       * Rovers and Farquhar 1973; Fungaroli and Steiner 1979.

(1) active methanogenesis substantially reduces leachate organic strength (by decomposing organic acids and alcohols) and increases pH; and (2) vigorous methanogenesis does not always occur in landfills because the landfill environment is much less than optimum for the methane bacteria (see Table 3).

Predicting leachate ocntaminant concentrations

The types, amounts, and production rates of contaminants appearing in the leachate at a landfill site are influenced by several factors: refuse type and composition; refuse density, pretreatment, placement sequence, and depth; moisture loading to refuse as influenced by the factors described for the WBM; temperature; and time.

Accurate quantification of these factors and their impact in the very heterogeneous conditions found in a landfill is difficult. The mechanisms and extent to which they influence contaminant release are poorly documented. It is therefore necessary to rely on data and experience from other landfill investigations and to apply them to landfills under study. Caution is necessary, since the conditions of the landfill investigation are often not documented and may not be the same as those being considered.

Several studies have been undertaken since the late 1960's in which investigators have collected leachates from refuse under various conditions and have measured their contaminant concentrations. Some have involved actual landfill sites, but the majority have been conducted with the use of lysimeters to simulate landfills. The lysimeters have ranged from small (volume <1 m³) laboratory units to large (volume >10 m³) field units. Lysimeters have often been preferred by investigators because it is easier to control conditions and make measurements of the leaching process with a lysimeter than with a full-scale landfill. Unfortunately, many lysimeter studies have not successfully simulated landfill conditions and have therefore produced atypical leachate. In comparison with the representative southern Ontario landfill conditions given in Table 3, lysimeters have often been too warm (+2O°C), too wet (several times greater than natural PERC), too shallow (<1 m), and too short in duration.

Lysimeter leachates have tended to be stronger than those experienced in the field, often because methanogenesis was not vigorous.

Leachate contaminant production curves
Lu et al. (1985) have produced an extensive review of investigations reporting leachate production and contaminant concentrations. They have combined the data obtained from these studies to produce contaminant production curves similar to those shown in Fig. 5. Plots for BOD⁵, iron (Fe), chloride (Cl), and ammonia nitrogen (NH₃-N) have been reproduced in

        Fig. 6. Leachate contaminant concentration variation with refuse age (from Lu et al. 1985).

Fig. 6. As might be expected, the data exhibit substantial scatter due to the wide range of conditions under which the studies were performed. In particular, some studies made use of atypical lysimeters and generally produced the higher concentrations shown. Consequently, the plots and models produced by Lu et al. (1985) represent upper limits for leachate contaminant concentrations at field installations.

Some investigators (Fungaroli and Steiner (1979), Ham (1980), Wigh and Brunner (1981), and McGinley and Kmet (1984)) have produced data which appear to reflect field conditions more closely. Their work has also been conducted and reported in such a way that the impact of some of the important factors such as compacted density, moisture addition, depth, and refuse age can be evaluated. These data sources are therefore particularly important in the prediction of leachate contaminant concentrations.

McGinley and Kmet (1984) have attempted to combine the data from these more realistic studies to produce leachate contaminant production curves. Examples of their plots of several data sets are shown in Fig. 7. One data set has been eliminated from each of the upper and lower graphs in Fig. 7. (These appeared to be clearly different from the others.) The graphs show leachate chemical oxygen demand (COD) as a function of moisture loading to the refuse in units of litters (L) of leachate per kilogram (kg) dry refuse (as opposed to time) in an attempt to normalize the data. The upper graph expresses COD as mg.m.L^-1 in the leachate while the Iower one uses mg COD leached per kg dry refuse. While the data are scattered there is a reasonably good trend shown in each case. Young leachates exhibit CODs in the range of 30 000 to 50 000 mg.L^-1, while leachates from old, extensively leached refuse have CODs generally less than 2000 mg. L^-1.

Similar plots for other leachate contaminants were prepared by McGinley and Kmet (1984). Most data tend to level off at moisture Ioadings of 5 L.kg^-1, indicating that the limit of leachate contaminants has been reached.

 Fig. 7. Leachate COD production curves (adapted from McGinley and Kmet 1984)

Fig. 8. Leachate chloride production curve (adapted from McGinley and Kmet 1984).

Method for predicting teachate contaminant concentration
Figure 8 (adapted from the chloride (Cl) graphs of McGinley and Kmet) presents information for developing a rational method to calculate leachate contaminant concentrations in a field situation. It shows four cells of a landfill, A to D, one above another. The geometry and the dry refuse mass density are the same for all cells in this simplified illustration. However, each cell has been placed at a different time and thus has a different age and leaching history; the latter is given as:

PARQUHAR

Table 4. Example to calculate leachate chloride concentration
                

A WBM analysis has been performed and has predicted that the leachate percolation at this time is 0.1 m/month or 0.6 m for the 6-month time interval used in this analysis. This liquid will flow through all four cells in sequence, picking up contaminants as it goes. Thus each cell will move along the contaminant production curve by an amount equal to