Fuel Processing Technology 81 (2003) 223 246 www.elsevier.

com/locate/fuproc

Biomass leachate treatment by reverse osmosis

B.M. Jenkins a,*, J.D. Mannapperuma b, R.R. Bakker c
a

Department of Biological and Agricultural Engineering, University of California, 1 Shields Avenue, Davis, CA 95616-5294, USA b California Institute for Food and Agricultural Research, University of California, Davis, CA, USA c International Rice Research Institute, Los Banos, Laguna, Philippines Received 1 October 2001; received in revised form 1 December 2002; accepted 1 December 2002

Abstract Leaching of alkali metals and chloride from biomass reduces fouling and slagging in furnaces and boilers, especially for grasses and other herbaceous crops used as fuel. Leaching also produces large volumes of leachate for disposal or treatment. Leachates obtained by water-washing rice straw were treated by reverse osmosis (RO) for water recovery and to evaluate ion and chemical oxygen demand (COD) rejection by the membrane as a function of feed concentration. Recovery above 90% was achieved with rejection of COD and ions other than carbonate also above 90%. Similar results would be expected to apply to other biomass leachates. Flux approaching a recommended 40 l m 2 h 1 was maintained at 90% recovery at a pressure half the maximum recommended pressure of the membrane tested. Flux declined above 90% recovery due to increasing osmotic pressure and probable fouling as salt concentrations likely exceeded solubility limits at the membrane surface. Cost of RO processing for a leaching facility handling 50,000 Mg year 1 of biomass and generating 1.9 Ml day 1 of leachate is estimated at $671 Ml 1 of leachate exclusive of brine disposal. The cost increases to $719 Ml 1 leachate for a facility of approximately half this capacity. RO costs add $7 8 Mg 1 to the cost of fuel processed. RO treatment reduces the volume of fresh water required for leaching and the volume of brine disposed. Where sufficient municipal wastewater treatment capacity exists the addition of RO to a leaching system may not be economically justified if costs assessed on the basis of strength are higher than costs assessed for flow. For treatment by evaporation ponds, RO offers a clear advantage over direct leachate disposal, reducing costs by 50% to 70%. Land application of both leachate and concentrate can take advantage of nutrients extracted from the biomass and could be used in the production of additional biomass. Application over an area one-tenth or more the original source area may be required to manage salts and to ensure that sustainable application rates are not exceeded. D 2003 Elsevier Science B.V. All rights reserved.
Keywords: Biomass; Straw; Fuel; Leaching; Reverse osmosis; Potassium; Chloride; Water recovery; Economics

* Corresponding author. Tel.: +1-530-752-1422; fax: +1-530-752-2640. E-mail address: bmjenkins@ucdavis.edu (B.M. Jenkins). 0378-3820/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0378-3820(03)00010-9

224

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

1. Introduction Water washing or leaching of biomass, especially of straws and other grasses, has been demonstrated to reduce fouling and slagging in furnaces and to improve combustion properties by removing large fractions of alkali metals and chloride [1 5]. The presence of alkali species in biomass reduces the melting temperatures of ash and contributes to catalyst deactivation in selective catalytic reduction (SCR) systems used to reduce NOx emissions. Chlorine is a combustion inhibitor, accelerates corrosion, enhances the volatilization of alkali, and is involved in the formation of acid gases and toxic emissions. The removal of these constituents from fuel is highly desirable. Leaching typically removes 80% or more of potassium, and better than 90% of Cl. Certain other elements can also be partially extracted by leaching. Leaching of biomass fuels might be incorporated into feedstock processing operations along the lines illustrated in Fig. 1. In some instances the leaching can be carried out in the field by leaving the biomass exposed to precipitation and then harvesting once dry. This technique, although generally inexpensive, is at best poorly controlled and suffers from weather variability and the risk of contamination of fuel with soil, organic matter loss and degradation of fuel quality, and unpredictable scheduling of cultural operations for subsequent crops. It does, however, provide for direct recycling of nutrients to the field, particularly potassium. An alternative is to leach biomass in controlled operations at the industrial site. In the latter case, the fuel quality can be improved relative to field leaching, but fuel dewatering and drying become more difficult and expensive, and a large volume of spent wash water or leachate is produced. Where a biomass facility is suitably located, leachate may be disposed of

Fig. 1. Leaching process options.

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

225

through land application, thereby returning valuable nutrients to the field. Leachate disposal by this method is likely to be regulated and carefully controlled. Without concentration of leachate and recovery of water, on-site leaching will considerably increase facility water consumption and waste disposal costs. Concentration of leachate offers the potential to recycle water to the process and recover materials or reduce the cost of disposal. In comparison with distillation and other desalination and concentration techniques, reverse osmosis (RO) has emerged as one of the most economical for feed water concentrations up to 50,000 ppm total dissolved solids (TDS) with product water typically below 500 ppm TDS [6]. Energy consumption is also typically lower than distillation and other methods. To evaluate the potential for RO concentration of biomass leachates, two trials were conducted using brines obtained by leaching rice straw. Leachate was concentrated using a pilot plant membrane testing unit. Composition and other characteristics of the water passing through the membrane (permeate) and the concentrated brine retained (retentate) were determined and used along with pilot performance in a preliminary economic assessment of RO processing at larger scales.

2. Methods 2.1. Leaching Whole rice straw was leached in two 10-kg batches using air-dried material (9.4 kg dry basis). Straw was obtained in bales from commercial fields of M-202, a medium grain Japonica variety grown in California. Straw was removed from the bales as loose material prior to leaching. Both batches were soaked fully submerged in deionized water for 24 h at room temperature without agitation. The first batch was soaked in 325 l of water, the second in 350 l. Leachate was drained from each batch and used separately in the two RO experiments. Straw from the first batch was subsequently soaked another 24 h in 378-l deionized water to evaluate the extent of leaching occurring during the first soak. Eightythree percent of K was leached in the first soak, increasing to 90% for the second. Cl removal was 88 96% after the first soak. Leachate from the second soak was not used in any of the RO experiments, and is not discussed further. 2.2. Reverse osmosis pilot plant and RO module A single-stage pilot plant membrane testing unit was used in the two experiments. The pilot plant consisted of a 100-l feed tank, Wanner D10 piston pump with variable frequency drive, pressure vessel, throttle valve, inlet and outlet pressure gages and stainless steel connecting piping, as illustrated schematically in Fig. 2. The system had the flexibility to accommodate several standard pressure vessels. These experiments used a pressure vessel for 64-mm (2.5 in.) diameter by 1-m (40 in.) long membrane modules. A model SH2540 spiral membrane module manufactured by Desalination Systems was used [7]. This module incorporates a low-pressure thin film composite membrane intended for brackish water demineralization. The performance characteristics as reported by the

226

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

Fig. 2. Pilot plant RO testing unit.

manufacturer for this module using a municipal water are listed in Table 1. The results in Table 1 were obtained at 25% water recovery (based on fraction of initial feed volume recovered as permeate) and pH of 7.8 in the feed. Total module surface area was 2.42 m2. The flow rate of the module was reported as 95 1 l h at 1.1 MPa pressure and 25 jC when operating at 25% recovery with a 1000 mg l 1 solution of NaCl. This translates to a permeate flux of 40 l m 2 h 1. The reported average salt rejection under these conditions was 96%. 2.3. RO trials The first RO trial (designated RO-1) was conducted in batch concentration mode starting with 100 l of leachate from the first straw batch. Leachate was cleaned up through a 5-Am prefilter when pumped from the leaching tank to the feed tank. Flow rate along the

Table 1 Demineralization characteristics of desalination systems SH2540 membrane [7] Item Sodium Calcium Magnesium Potassium Bicarbonate Sulfate Chloride Silica Total dissolved solids Concentration (mg l 1) 68.0 80.0 21.0 4.1 132.0 163.0 51.0 9.3 528.0 Rejection (%) 91.4 99.6 99.3 90.0 94.0 99.7 91.0 87.0 96.1

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

227

membrane module was maintained at 22 l min 1 corresponding to the maximum capacity of the feed pump. Inlet pressure was maintained at 2 MPa, and the pressure at the outlet remained stable also at 2 MPa. Six sets of readings were collected at six water recovery levels corresponding to progressive draining of 20-l permeate between readings (total feed volume was reduced by 20 l for each sample). Both permeate and retentate were returned to the feed tank for the first 5 min to stabilize concentrations and allow steady operation at each recovery level before sampling. At each increment, permeate flow rate was measured along with electrical conductivity and pH. Samples of permeate and retentate were collected for analysis of ions, acids, sugars and COD. Concentrations of ions were determined on a Waters HPLC using a model 430 conductivity detector and WISP sample processor. Cations were analyzed on a 4.6 50 mm Waters IC-PAK cation M/D column. Anions were analyzed on a Waters IC-PAK anion column of the same size. Organic acids were analyzed on the same HPLC using a Supelco C610H 7.8 mm 30 cm column and UV detector. Sugars were analyzed on a J&W Scientific Accubond amino 5-Am column with a differential refractometer detector. The second RO trial (RO-2) was also conducted in batch concentration mode but using 250 l of leachate. Because the feed tank volume was only 100 l, additional feed was pumped to the feed tank to maintain 100-l volume until the total feed volume was accommodated. All other operating procedures were the same as the first trial, except that retentate temperature was also measured. Flux values were adjusted to standard conditions (25 jC) to correct for the changing water viscosity according to [6]: Jo J 1:03T 25 1

in which J is the measured flux, Jo is the flux corrected to 25 jC, and T is the feed temperature (jC). As temperature was not measured during the RO-1 trial, flux values were not corrected and so are higher on average by a factor of approximately 1.3 compared with those reported for RO-2.

3. Results 3.1. Leaching effectiveness Data in Table 2 show that the leaching treatment was of typical effectiveness in removing K and Cl. Other elements with high to moderate extractions were Na, P, S, Mg and, to a lesser extent, Ca. Total ash was reduced by 15% to 16%. Ion concentrations in the two leachates are listed in Tables 3 and 4. Material balance calculations indicate that element recoveries are for the most part excessive, as shown by Fig. 3, with RO-2 yielding higher deviations from closure than RO-1. For these balances, the dry weight of leached straw was derived by silica balance, assuming silica to be insoluble. Silica in leachate was not analyzed. Ash recoveries are as expected, given the soluble fraction from previous determinations [1], but potassium recovery is high. Ion concentrations for RO-2 leachate are generally higher than those for RO-1 leachate. This is partly consistent with the higher

228

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

Table 2 Rice straw compositions Treatment Higher heating value (MJ kg

1

Untreated ) 14.75

RO-1 15.21

RO-2 15.15

Ultimate analysis (%dry matter) Carbon Hydrogen Nitrogen Sulfur Ash Chlorine Oxygen Ash analysis (%ash) SiO2 Al2O3 TiO2 Fe2O3 CaO MgO Na2O K2O P2O5 SO3 Cl CO2 Undetermined Ash in fuel (%dry matter) SiO2 Al2O3 TiO2 Fe2O3 CaO MgO Na2O K2O P2O5 SO3 Cl CO2 Undetermined

37.87 4.61 0.63 0.14 20.87 1.01 34.87

39.75 4.74 0.6 0.1 17.44 0.12 37.25

39.53 4.76 0.53 0.08 17.89 0.04 37.17

72.28 0.07 0.02 0.26 2.08 2.06 2.71 11.8 1.78 1.11 4.07 0.1 1.66

86.53 0.43 0.02 0.29 2.15 1.26 0.5 2.45 0.66 0.51 0.07 0.14 4.99

89.12 1.19 0.05 0.43 2.47 1.35 0.32 1.39 0.49 0.41 < 0.01 0.12 2.66

15.08 0.01 < 0.01 0.05 0.43 0.43 0.57 2.46 0.37 0.23 0.85 0.02 0.35

15.09 0.07 < 0.01 0.05 0.37 0.22 0.09 0.43 0.12 0.09 0.01 0.02 0.87

15.94 0.21 0.01 0.08 0.44 0.24 0.06 0.25 0.09 0.07 < 0.01 0.02 0.48

extractions of soluble elements for RO-2 compared with RO-1 (Table 2), but likely indicates error in the HPLC determinations for RO-2. The very high deviation for alumina is due to the low and probably erroneous determination for the untreated straw ash (Table 2), but alumina is in any case present only at very low concentrations. Concentrations of ions in the deionized water were very low and do not account for the large closure errors observed for some elements.

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

229

Table 3 Concentrations (mg l 1) in leachate (L), permeate (P) at six different recoveries, and final retentate (R), RO-1 Sample L P0 P20 P40 P60 P80 P95 Ra Na 112 4 4 4 6 12 53 939 K 989 32 37 43 60 122 527 9160 Mg 42 Ca 2 2 2 CO3 186 22 22 10 25 54 201 Cl 288 9 11 13 24 54 238 3410 PO4 95 SO4 86 COD 1660 117 93 86 72 105 261 15,500 pH 6 6 6 6 6 6 6 7

601

839

Blank indicates not detected. Subscripts on permeate (P) labels indicate percent water recovery for the sample. a Retentate collected at 95% recovery.

Solid liquid distributions between straw and leachate for the main soluble elements are shown in Figs. 4 and 5. The high extraction levels for K, Na and Cl are apparent. About one-third to one-half of sulfur and magnesium and 15 30% of phosphorous are extracted into the leachate. Calcium is mostly insoluble. 3.2. RO trials 3.2.1. RO-1 Permeate flux (uncorrected) at the beginning of the first trial at zero recovery was 56 l m 2 h 1, leveling at around 50 l m 2 h 1 between 20% and 80% recovery before declining to 32 l m 2 h 1 at the last reading around 95% recovery (Fig. 6). Electrical

Table 4 Concentrations (mg L 1) in leachate (L), permeate (P) and retentate (R) at six different permeate recoveries, RO2 Sample L P1 P2 P3 P4 P5 P6 R2 R3 R4 R5 R6 Na 172 6 8 8 9 13 429 282 323 435 721 2021 K 1538 54 72 81 98 160 3650 2592 2981 4209 6448 11,122 102 124 225 363 754 Mg 94 Ca 34 3 1 0 0 1 1 13 17 73 124 215 CO3 303 81 48 36 16 50 290 376 325 190 183 2293 Cl 404 33 19 23 29 61 7964 748 1094 1161 1443 6290 197 254 306 402 2325 228 259 289 432 1598 PO4 195 SO4 138 COD 1300 21 30 29 41 42 42 1500 2400 2700 6000 35,800 pH 6 5 5 5 5 5 8 6 6 6 7 7

Blank indicates not detected. Same subscripts on P and R labels indicate samples collected at same time. Actual permeate recovery not specifically measured at time of sampling due to leach-tank to feed-tank transfer during processing of first 150 l.

230

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

Fig. 3. Closure errors for material balances (excess recovery shown as positive value).

conductivity (EC) of permeate was 95 AS cm 1 at the beginning of the trial when the EC of retentate was 2450 AS cm 1. This translates to a 96.1% rejection of EC by the membrane. Electrical conductivity of the permeate increased gradually up to about 80% recovery, and then increased sharply to 1350 AS cm 1 at the end when EC of retentate was 21,500 AS cm 1 (Fig. 6). EC rejection at the end was 93.7%. Ion concentrations, chemical oxygen demand (COD) and pH are listed in Table 3 for leachate, permeate at six recovery levels, and the final retentate at 95% recovery. COD of permeate varied between 72 and 261 mg l 1, increasing at recoveries above 60%. COD of retentate increased from an initial 1660 (in leachate) to 15,500 mg l 1 at the highest

Fig. 4. Solid liquid distribution of soluble elements from RO-1.

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

231

Fig. 5. Solid liquid distribution of soluble elements from RO-2.

recovery. Permeate pH increased slightly towards the end of the trial. Retentate pH increased from 6.1 in leachate to 6.7 at the end. Final ion rejection values were above 90% except for carbonates. Oxalic, citric, acetic, propionic, succinic, formic and fumaric acids were identified in the leachate and retentate, with the first four having the highest concentrations ranging up to 5000 mg l 1 for oxalic acid in the final retentate. Glucose (67 mg l 1 in retentate) and sucrose (168 mg l 1 in leachate) were detected but not consistently in all samples and further analysis would be required to assess sugar concentrations. 3.2.2. RO-2 Permeate flux at the beginning of the second trial was 41 l m 2 h 1. Flux declined in a manner similar to the first trial up to a retentate EC of 13,500 AS cm 1. At the end of the test with retentate EC at 50,000 AS cm 1, the flux dropped to less than 4 l m 2 h 1. Temperature slowly increased from 28 to 38 jC through most of the run, increasing rapidly to above 50 jC at the end. The leakage characteristics of the

Fig. 6. Permeate flux (uncorrected) and electrical conductivity, RO-1.

232

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

membrane were similar to the first trial except at the last reading. The exceptional behavior at the end is a result of high osmotic pressure and also of fouling. Fig. 7 shows the variation of permeate flux and EC as functions of the EC of the retentate (data from last reading not shown). Fig. 8 shows osmotic pressures calculated from ion concentrations in the feed. The pressure differential during RO-2 declines from 1.8 MPa at the start to 0.3 MPa at the end. For these leachates, the osmotic pressure (p, MPa) can be estimated from feed EC (AS cm 1) as: p 4:65 1010 EC2 5:73 105 EC r2 0:9991 2

Fig. 8 also shows the total ion concentration as a function of the EC. For both osmotic pressure and ion concentration, the regressions have been performed using only the RO-2 data. The RO-1 values (two points only for osmotic pressure, L and R6) correspond well to the RO-2 results. At lower values of EC, the concentrations are close to those predicted by the more commonly applied linear conversion: c = 0.64EC. Ion concentrations, COD and pH for RO-2 are listed in Table 4 for leachate as well as permeates and retentates at six different observations (retentate for the first observation is the leachate). Retention remained above 90% for ions other than carbonate except for the last observation at high retentate EC. The concentrations at the last point may exceed solubility limits of the salts at the membrane surface. Leakage of COD during RO-2 remained extremely low throughout the trial. The COD in the permeate reached a maximum of 42 mg l 1 when the retentate COD was 35,800 mg l 1. COD rejection remained in the range of 98 99% (Fig. 9). This is somewhat higher than the 93% and 98% COD rejection seen in the first trial at the first and last observations. Permeate pH remained at about 5.0 until the end when it increased to 8.3. Retentate pH increased from 6.2 (leachate) to 7.0 at the end. A similar suite of acids and sugars was detected in the leachate, retentate and permeate, with acetic, oxalic, citric and propionic again having the highest concentrations of the acids. Oxalic acid was found at 14,500 mg l 1 and acetic acid at 24,000 mg l 1 in the

Fig. 7. Permeate flux (corrected to 25 jC) and electrical conductivity, RO-2.

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

233

Fig. 8. (a) Ion concentrations (c, mg l 1) and estimated osmotic pressures (p, MPa) from feed compositions. (b) Osmotic pressure variation with total ion concentration (wt.%). Regression equations for (a): p = 4.650 10 10EC2 + 5.730 10 5EC, r2 = 0.9991; c = 4.922 10 6EC2 + 7.783 10 1EC, r2 = 0.9898.

final retentate. Acid rejection remained above 95% even at the last observation. Glucose and sucrose were identified only in the final retentate and permeate for RO-2. Glucose concentration was greatly elevated (10,990 mg l 1) in final retentate compared with RO-1, but there is high uncertainty in these values. For both RO-1 and RO-2, carbonate rejection is mostly below 90% and extends to as low as 73%. Equilibrium fractions computed for the range of pH observed in both

234

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

Fig. 9. Permeate COD and COD rejection for RO-2.

permeates and retentates are shown in Fig. 10. The individual equilibrium species concentrations were computed from total carbonate concentrations, pH, and carbonic acid and bicarbonate ionization constants, K1 = 4.3 10 7 and K2 = 4.7 10 11 [8], as: H2 CO3 Ct K1 K1 K2 1 2 H H Ct H K2 1 K1 H 3

HCO 3

Fig. 10. Equilibrium fractions of carbonic acid and bicarbonate computed for the range of pH observed in permeates and retentates, RO-1 and RO-2. Fraction of CO3= is not shown and is very low for this pH range.

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

235

CO 3

Ct H H 2 1 K2 K1 K2

For the pH range observed, [CO3=] is near zero. pH of retentates is mostly between 6 and 7, with permeates between 5 and 6. Within these ranges, a large fraction of carbonate exists as carbonic acid, which may explain the low rejection of carbonate by the membrane. 3.3. General performance The appearances of the leachates, permeates and retentates from the two experiments are shown in Fig. 11. Permeates remain clear throughout. Permeate fluxes at 2 MPa pressure are close to the maximum recommended level of 40 l m 2 h 1. There is ample margin left to the maximum pressure of about 4 MPa for this class of membrane modules. Recovery levels up to 90 95% may be reasonably anticipated. Calcium carbonate and possibly calcium sulfate scaling could become problems at higher recovery levels, in which case additional pretreatment, a different type of module or the use of seeded RO might be called for [6,9]. With high silica materials such as rice straw, silica in leachate could also lead to scaling, but no leachate silica measurements were made. Silica in excess of 120 ppm is a distinct problem, as metal hydroxides can absorb or complex it and catalyze the precipitation. Further work is required to determine the potential significance of this problem.

4. Economic analysis The costs of treating biomass leachate by RO were adapted from an earlier study of Bakker [10] based on the performance results described here and the cost estimating method of Ray [11]. Costs are estimated from assumptions of the operating recovery level for the RO unit, the amount of water required for leaching the biomass, and the size, efficiency and availability of the power plant using the leached biomass fuel. Three different cases are evaluated for the size and efficiency of power plant to examine the effect of scale on RO plant capital and total costs. The size of RO facility considered is based on the assumption that only 25% of the fuel used by the biomass power plant would need leaching. This might be the case, for example, of a wood-fired power plant accepting

Fig. 11. Appearance of leachate (L), permeates (P), and retentates (R) from RO experiments.

236

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

straw fuel for blending. To reduce fouling and slagging in the furnace, the straw could be leached prior to blending with the wood. Large-scale tests of blending leached rice straw with wood in this manner have been conducted [4]. The blending level is here assumed to be 25% of heat input in fuel. None of these assumptions is critical to the design of the RO unit; instead, they serve only to define the capacities of the units analyzed. Water use assumptions are listed in Table 5. Untreated biomass is assumed to be received at 10% moisture content, wet basis. Leaching is conducted at a net water application rate of 12 l kg-1 dry matter, a value arrived at through laboratory studies [9] and thought to yield adequate extraction of alkali and chloride. The net applied water is a combination of permeate recovered from the RO unit and fresh makeup water. Some direct recycling of water could also be employed in the leaching system, but only the net application is required for this analysis. Following leaching, the biomass is mechanically dewatered to 50% moisture and fired to the boiler. The 50% moisture is based on preliminary results obtained with leached rice and wheat straw in a commercial roller press, and is typical of moistures in sugar cane bagasse following juice expression. Of the original dry matter fed to the leaching system, a part is taken into solution in the leachate. Dry matter recovery is assumed to be 95%. Of the applied leaching water, approximately 0.8 l kg 1 is carried into the boiler as fuel moisture (or if the fuel is first further dried, moisture is lost to the atmosphere), leaving 11.2 l kg 1 as leachate to be processed. A permeate recovery of 80% is assumed for the RO unit, based on results presented above (see, for example, Fig. 6). Operating at this recovery returns approximately 9 l kg 1 for leaching, with a little over 2 l kg 1 of concentrated brine requiring disposal or other processing. The fresh makeup water requirement is about 3 l kg 1. The three cases analyzed are (1) a 25 MWe power plant operating at 20% net station electrical efficiency and typical of current commercial technology, (2) a 25 MWe power plant operating at 30% efficiency as might occur in the future, and (3) a 10 MWe power plant operating at 20% efficiency, also of current conventional design. The plant availability in all cases is 80%, and the RO system is assumed to operate with the same availability. Use of waste heat or water in cogeneration applications is not considered.

Table 5 Water use assumptions for economic analysis Moisture content of untreated biomass (% wb) Moisture content of dewatered leached biomass (% wb) Fraction of biomass recovered after leaching (%) RO operating recovery (%) Water remaining as moisture in biomass (l kg 1 dm)a Makeup fresh water (l kg 1 dm) Water recovered from leachate (permeate) (l kg 1 dm) Water applied to biomass for leaching (l kg 1 dm) Leachate (l kg 1 dm) Brine for disposal or other recovery (l kg 1 dm)
a

10 50 95 80 0.8 3.1 8.9 12 11.2 2.2

dm = dry matter. Totals may not add due to rounding.

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

237

The costs of processing leachate are shown in Table 6. The capacities of the RO plants to process leachate are 1.9, 1.3 and 0.8 Ml day 1 for the three cases. Capital costs ($ Ml 1 day of capacity) are assumed to scale with capacity, M, at a scale factor, s = 0.87, from a reference plant size Mo = 3.8 Ml day 1 [11]. Capital costs for each case were scaled by the ratios listed in the table, defined by (M/Mo)s 1. Capital costs included were site development costs, utilities (exclusive of electric power, makeup water and brine disposal) and equipment. Site development costs may be conservative, as existing power plants retrofitting RO equipment would not need the level of development considered. Land costs are not included under the assumption that these would be otherwise allocated to the power plant. Indirect capital costs include construction overhead costs and contingency fees. Total capital costs per unit daily capacity were amortized over the assumed 10 year life of the facility using an interest rate of 12% year 1.
Table 6 Costs of processing biomass leachate by RO Power plant size (MWe) Net station efficiency (%) Power plant availability (%) Blended fuel fraction (% HV)a Leaching capacity (Mg day 1 dm) RO plant capacity (Ml day 1) Capital cost scale ratio (s = 0.87) Capital costs ($ Ml 1 day) Direct: 25 20 80 25 168.8 1.9 1.09 25 30 80 25 112.5 1.3 1.15 10 20 80 25 67.5 0.8 1.23

Indirect:

Site development Utilities Equipment Total direct Construction Contingency Total indirect costs

Total capital costs: Capital recovery costs ($ Ml 1 at 12%/10 years)

28,929 31,822 448,405 509,157 63,645 54,966 118,610 627,767

30,495 33,545 472,675 536,715 67,089 57,941 125,030 661,745

32,589 35,848 505,130 573,567 71,696 61,919 133,615 707,182

380 Operation and maintenance costs ($ Ml 1) Energy Membrane replacement Labor and overhead Mechanical supplies Chemicals Filters Total O&M Total costs ($ Ml 1) Total daily cost ($ day 1) Total daily energy revenue ($ day 1 at $0.075 kWh 1) Cost/revenue (%)
a

401

429

114 48 53 24 40 13 445 671 1264 45,000 2.8

114 48 53 24 40 13 445 692 869 45,000 1.9

114 48 53 24 40 13 445 719 542 18,000 3

Heating value. Totals may not add due to rounding.

238

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

Operation and maintenance costs include energy in pumping and other power requirements, membrane replacement, labor and overhead, mechanical supplies, chemicals and filters. The operating costs do not include brine disposal or purchase of makeup water. These are handled separately in a comparison of the leaching system with and without RO treatment of leachate. Total RO treatment costs for the three cases are $671, $692, and $719 Ml 1 of leachate processed. The costs are similar to other costs for RO treatment of brackish waters [6]. The cost per unit leachate processed by RO for case 2 is higher than case 1 because of the lower fuel requirement and smaller RO unit needed for this higher efficiency plant. The total daily cost of operation is, however, nearly $400 lower than for the less efficient power plant of the same size because of the smaller volume of leachate. The cost of RO for leachate treatment would be about 3% of the energy revenue for conventional power plants operating with prices of $0.075 kWh 1, declining to 2% for more efficient plants (case 2). The overall feasibility of leaching includes costs other than RO treatment of leachate alone. Bakker [10] estimated the cost of RO treatment at about $6 Mg 1 dry matter, representing about a third of the total cost of on-site leaching. The RO costs computed here are higher at $7 8 Mg 1 dry matter and add upwards of 45% to the cost of leaching. The total added cost of leaching generally makes the process economically unattractive at current electricity prices and power plant efficiencies unless specific subsidies or other incentives such as tipping fees are applied to reduce net fuel costs. For conventional biomass power plants, fuel costs below about $10 Mg 1, exclusive of leaching costs, would be required to justify leaching for electricity prices in the range of $0.05 kWh 1. Increasing the net station efficiency using more advanced conversion technology would permit leaching at higher delivered fuel costs. For less expensive fuels or fuels that are subject to incentives but which are not used due to problems with slagging, fouling, acid gas emissions or catalyst deactivation, leaching may still be an effective part the plant operation with the inclusion of RO depending on the local cost of fresh water and brine disposal. A leaching system could be deployed with or without an RO treatment system. In the case without RO, the cost of purchasing or otherwise supplying the fresh water for leaching along with the cost of full leachate disposal should be less than the cost of the system with RO added. The effect of RO is to reduce the volume of fresh makeup water needed for leaching and the volume of brine requiring disposal. The sensitivity of the leaching system cost with and without RO to the makeup water cost and the brine disposal cost is shown in Fig. 12 for case 1, the facility generating leachate at a rate of 1.9 Ml day 1. The total cost shown is without the cost of the leaching system itself, as this is assumed to be the same regardless of whether an RO unit is included or not. Without RO, the total cost shown is just the sum of the water acquisition and brine disposal costs. With RO, the cost also includes the cost of RO treatment for this scale, $671 Ml 1 (Table 6). The total cost is per unit volume of leachate generated. Brine and water costs are per unit volume of brine and water, respectively. Fig. 12 shows that the system with RO is much less sensitive to brine disposal cost than the system without it. Breakeven for the two systems occurs at the intersection for equal water costs. As the makeup water cost increases, breakeven occurs at lower brine disposal

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

239

Fig. 12. Cost ($ Ml 1 of leachate less cost of leaching system) with (solid lines) and without (dashed lines) RO treatment, 1.9 Ml day 1 leachate generation.

cost. For the assumptions of 80% recovery and no brine disposal cost, the RO system is preferred when the water acquisition cost exceeds $840 Ml 1. As the RO recovery changes, this cost changes as shown in Fig. 13. Brine produced by the RO unit is of higher strength than the initial leachate and cost to dispose of concentrated brine will in general be higher per unit volume disposed. The breakeven cost of concentrated brine disposal, Eq. (6), is derived from the costs of

Fig. 13. Makeup water costs ($ Ml 1) above which RO is preferred when brine is disposed of at no cost, 1.9 Ml day 1 leachate generation.

240

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

unconcentrated brine (leachate) disposal, makeup water and other operating costs of the RO system:   1 Cc;RO 6 Cb;RO Cb;L fCw 1f Vb;RO in which Cb,RO is the breakeven cost for concentrated brine disposal ($ Ml 1); Cb,L is the disposal cost for unconcentrated leachate ($ Ml 1); Cw is the cost of makeup water ($ Ml 1); Cc,RO is the daily operating cost of the RO system ($ day 1); Vb,RO is the daily volume of concentrated brine from RO (Ml day 1); f is the RO permeate recovery fraction. Table 7 summarizes the breakeven conditions for RO compared with direct leachate disposal (or use). If the makeup water is available at no cost, breakeven occurs when the leachate is disposed for $840 Ml 1. The RO system generates savings for any concentrate disposal cost at or below this value. If the leachate disposal cost increases to $1000 Ml 1, the concentrate disposal cost can increase to as high as $1644 Ml 1 before the RO system becomes uncompetitive. With makeup water at $50 Ml 1, breakeven occurs at $640 Ml 1, and for makeup water at $500 Ml 1, breakeven is at $340 Ml 1. With water and leachate disposal both at $500 Ml 1, concentrated brine disposal could rise to $1140 Ml 1 before RO became uncompetitive. The latter costs would add roughly $11.50 Mg 1 (dry matter) to the cost of leached fuel, exclusive of the other leaching system costs. Comparative costs for Sacramento Valley communities within the primary California rice growing region, as adapted from Ref. [12], suggest that makeup water, if purchased from municipal suppliers, could range up to about $330 Ml 1, with leachate disposed as municipal wastewater costing $600 900 Ml 1. At 80% recovery, disposal of concentrated brine from RO could cost $2500 to 4200 Ml 1. At the highest water and leachate disposal costs, breakeven costs for concentrated brine disposal are $2500 Ml 1, less than the projected cost of disposal. With no (or very low) cost for makeup water and the lowest cost for leachate disposal, RO is not competitive unless use for concentrate becomes available giving it a value of around $400 Ml 1. This analysis assumes that existing wastewater treatment facilities would be able to accept the volumes produced by a leaching system, and ignores any transportation cost. Where a power plant or leaching facility is permitted for zero discharge of wastewater, the use of RO would need to be compared against the land and other costs for holding and
Table 7 Breakeven concentrate disposal costs for RO system with various makeup water and unconcentrated leachate disposal costs Make-up water cost ($ Ml 1) 0 0 50 500 500 Leachate disposal cost without RO ($ Ml 1) 840 1000 640 340 500 Breakeven concentrate disposal cost with RO ($ Ml 1) 840 1644 640 340 1140

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

241

Fig. 14. Area, annual cost, and cost per unit leachate for evaporation pond operating at 1 m net evaporation per year, 10-year life, 12% interest rate, leachate generation rate of 1.9 Ml year 1. Vertical lines indicate annual volume to be handled by disposal system with and without RO.

evaporation ponds to handle a much larger volume of brine if no other use can be found for it. Total annual volume of leachate for a 1.9 Ml day 1 facility (case 1) is 550 Ml. A 55-ha pond would be required if this leachate were to be disposed of through the use of an evaporation pond with a net annual evaporation rate of 1 m. Using RO to recover water, the liquid volume for disposal would decrease to 110 Ml year 1, with a pond area of 11 ha. Area required and estimated cost for disposal via evaporation ponds are shown in Fig. 14. Conditions specific to a 1.9 Ml day 1 facility are summarized in Table 8. Costs are based on triple-lined ponds with assumed 10-year life and 12% annual interest rate on capital. A 55-ha pond to handle the full leachate disposal would have an annual capital charge under these conditions of $3.2 million with a unit cost of $5831 Ml 1 of leachate. The 11-ha pond to handle the concentrate from the RO system would have an annual
Table 8 Estimated costs of leachate and concentrate disposal in evaporation ponds, 1.9 Ml day 1 leachate production Leachate disposal without RO Disposal volume (Ml year 1) Pond area (ha) Capital cost ($ 106) Annual capital charge ($ 106 year 1)a Capital charge ($ Ml 1 leachate) RO cost ($ Ml 1 leachate) Solids disposal ($ Ml 1 leachate) Total cost ($ Ml 1 leachate)
a

Concentrate disposal with RO 110 11 3.6 0.6 1166 671 80 1917

550 55 18.1 3.2 5831 80 5911

Ml

Assumes 10-year life, 12% interest rate per year. For 30-year life and 5% interest rate, total costs are $2223 without RO, and $1180 Ml 1 with RO.

242

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

charge of $0.6 million, with a unit cost of $1166 Ml 1 of leachate (but still $5831 Ml 1 of brine). Extending pond life to 30 years and reducing interest rate to 5% year 1 would decrease capital charges to $2100 Ml 1 for leachate and $400 Ml 1 for concentrate evaporation, per unit of leachate produced. In either case, adding the RO system cost of $671 Ml 1 results in a lower total cost compared to direct leachate evaporation. Solids (salt) accumulation in the ponds, for the production levels of case 1, would amount to 2200 Mg year 1. Over a 10-year life, disposal of accumulated solids in a landfill with a tipping fee of $20 Mg 1 would add $440,000 or $80 Ml 1 leachate. Dehydrated solids might also be land-applied for their fertilizer value. Other alternatives for leachate or brine disposal include utilization in landfill bioreactors or digesters and land application as fertilizer. Landfills receiving liquid wastes could utilize brine in promoting decomposition of solid waste, although typical tipping fees of $20 30 Mg 1 ($20,000 $30,000 Ml 1) could make this infeasible. However, with concentrated brine, the annual tipping fee of $2.2 million for case 1 could be less than the annual cost for full leachate disposal by evaporation pond ($3.2 million at 12% interest and 10-year life) depending on life and financing conditions for the latter. A suitably sized landfill would need to be located in proximity to the plant (or the leaching facility located at the landfill site) to handle the volumes generated. Land application could be useful for the production of additional biomass fuel due to the nutrient value of extracted K, P and other constituents. For rice straw, the roughly 50,000 Mg year 1 of biomass leached in case 1 would originate from approximately 10,000 ha of land. Distribution of leachate or concentrate over an equal area would be necessary to maintain loading equal to current agronomic practice. Concentrated application may be possible, but sustainable operation would still likely require a reasonably large land area. The application rates for various constituents in leachate as a function of the land area available are shown in Fig. 15 (rates for concentrate would be slightly lower). As land area decreases, the frequency of application need not change, but the amount applied per unit area increases. The application rates for 10,000 ha are equal to those that occur with rain-washing in the field, and also if the straw is incorporated back into the soil. Similar rates for some of the species occur even with field burning for disposal, although much of Cl and S are volatilized. Of primary concern are the application rates of Na, Cl and COD as the available distribution area declines. Na concentration in rice straw leachate is about one-tenth that of potassium (as K2O equivalent). Cl concentration is about one-fifth that of K2O. At higher application rates, salinity could be a significant problem. Rice, for example, is a moderately salt-sensitive crop, exhibiting a 10% yield decrement for applied water EC (EC10) greater than 2600 AS cm 1 [13]. More salt tolerant crops, such as barley, cotton, sugar beets, wheat and date palm (4500 < EC10 < 6700 AS cm 1) or forage grasses such as tall wheat grass, rye grass and fescue (3900 < EC10 < 6,600 AS cm 1) could allow for concentrated application, but local conditions would dictate the feasibility of production. Some salt-tolerant tree crops could utilize leachate for irrigation water. Eucalyptus can tolerate up to about 10,000 ppm TDS [14] and might survive application of concentrates produced at 60% to 80% recovery, but biomass production would likely be heavily suppressed. Adjusted sodium adsorption ratios (adj. SAR) computed for leachate are in the range of 8 to 10, with Na and Cl concentrations in the range of 4 8 and 8 11 meq l 1,

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

243

Fig. 15. Application rates (kg ha 1 year 1) of leachate constituents, 1.9 Ml day 1 facility. Original source area of the leached biomass = 10,000 ha.

respectively. As irrigation water, leachate is categorized as moderately to severely restricted in quality and use [13]. For concentrates, adj. SAR is in the range of 20 100, and would be severely restricted in use without dilution. Sodium and chloride toxicity could occur in crops irrigated with either leachate or concentrate. Sodium toxicity is commonly observed in plants when leaf tissue concentration reaches 0.25 0.5% dry matter. Chlorine toxicity occurs with leaf concentrations in the range of 0.3 0.5%. For case 1 with a capacity of 1.9 Ml day 1 leachate, if all applied Na and Cl were accumulated uniformly throughout the plant tissues, an application area of 5000 to 12,000 ha would be required to avoid toxicity in a crop with a biological yield (dry matter) of 10 Mg ha 1 year 1. In this case, Cl would be the limiting ion. If all Na and Cl were concentrated in just the leaf tissues, the area required would increase by a factor of 3 to 20 depending on crop. These conditions are extreme, and application over smaller areas might still prove feasible. Fig. 16 shows the area required as a function of dilution ratio for leachate applied to rice. This analysis assumes a total evapotranspiration for the crop (ETc) of 1.2 m year 1, and a standing water level of 0.2 m in the paddies. Concentration of applied water in the field occurs through the crop evapo-transpiration. The field concentration ratio in this case is 1.2/0.2 = 6. If leachate were used as the sole source of water for rice, the production from case 1 would be sufficient to irrigate 46 ha. The field concentration by evapo-transpiration would lead to excessive salinity in paddy water, reaching concentrations above 17,000 mg l 1 TDS and 68 meq l 1 Cl. To meet the TDS limit (1600 mg l 1 or EC of 2600 AS cm 1) based on a 10% yield decrement, a dilution ratio of 16 is required assuming a TDS

244

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

Fig. 16. Area required as a function of dilution ratio for land application of leachate on rice. Dilution ratio to meet TDS concentration limit = 16.2 (744 ha). Dilution ratio to meet Cl concentration limit = 22.9 (1050 ha).

concentration of 100 mg l 1 in the dilution water (a value typical of Sacramento basin water available for rice irrigation). The application area to meet this constraint is around 750 ha. The more stringent Cl limit (Cl < 4 meq l 1 in paddy water) imposes a 23 dilution ratio for a Cl concentration of 0.17 meq l 1 in dilution water, achieved with leachate application on 1050 ha. The same areas (or slightly smaller) apply to RO concentrate application, but the dilution ratios would be higher by a factor of 5 for 80% recovery. These results suggest that leachate could be land-applied on an area roughly one-tenth that of the original source area without exceeding sustainable levels of salts, given sufficient dilution. Boron toxicity is also a problem in saline irrigation, with symptoms appearing at leaf concentrations of 250 to 300 ppm. Boron in leachate was not measured, but is typically below 10 ppm in rice straw [15] and thus is not expected to be a problem for leachates of the type tested. Even if all boron were leached from straw and taken up in crop tissues (again an extreme condition), toxicity should not occur if the application area were in excess of 200 ha. Odors due to fermentation of organic matter might prove to be a problem in land application if leachate or concentrate were stored prior to application. Other impacts also need to be considered, including effects on the market value of crops grown under saline irrigation due to changes in composition or other properties. Health implications aside, the burning (smoking) characteristics of tobacco, for example, a chloride-tolerant crop, become less desirable as the chloride concentration in the leaf increases [13], a phenomenon due to the inhibitory effects of chlorine on ignition and burning. The improvement in the ignition and burning characteristics of straw when Cl is leached involves the same mechanism [1]. K, P and S have economic value as fertilizer (nitrogen is not leached substantially and so does not contribute in this capacity). Based on commercial

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

245

fertilizer prices [16], the value of leachate might be as much as $680 Ml 1, and the value of concentrate as much as $3400 Ml 1 if salinity and COD can be properly managed. Land application of leachates as irrigation water and fertilizer for woody biomass crop production could have an advantage in the transfer of elements, particularly potassium, from high-fouling to low-fouling fuels, again assuming that salinity is properly managed. Subsequent use of the wood as fuel would pass these elements to the ash stream of the power plant (also typically land-applied) without the fouling and slagging associated with direct firing of the high-fouling materials (although care would be needed in fuel blending to avoid undesirable recombination of alkali and silica). Fuel and ash composition could be selectively controlled in this manner due to the selective nutrient uptake by different biomass species, with potential positive impacts on air emissions (reduced acid gas emissions) as well as the thermal characteristics of ash. Although such a strategy might be economically infeasible, leaching allows for some intriguing alternatives to conventional fuel management and production system design.

5. Conclusions Recovery of water from leachate by reverse osmosis can be expected to exceed 80%, and recoveries above 90% can be achieved. COD rejection and rejection of ions other than carbonate are above 90% for the membrane used when processing a potassium-chloriderich leachate from rice straw. Similar results would be expected to apply to other biomass leachates. Flux approaching a recommended 40 l m 2 h 1 was maintained to 90% recovery at a pressure half the maximum recommended pressure of the membrane. The large flux decline above 90% recovery observed in these experiments is due to the declining difference between applied and osmotic pressure with increasing concentration, and to possible fouling of the membrane as salt concentrations may have exceeded solubility limits at the membrane surface. Silica fouling due to processing leachate from high silica materials such as rice straw may also occur, but needs further investigation. The cost of RO processing for a leaching facility handling 50,000 Mg year 1 of biomass and generating 1.9 Ml day 1 of leachate is estimated at $671 Ml 1 exclusive of brine disposal. The cost increases to $719 Ml 1 for a facility of approximately half this capacity. RO adds $7 8 Mg 1 to the cost of fuel processed. Full leaching costs are likely two to three times this value, and depend additionally on the cost of fresh makeup water for leaching and on the cost of leachate or concentrated brine disposal. RO treatment of leachate reduces the volume of fresh water required for leaching and the volume of brine disposed. Where sufficient municipal wastewater treatment capacity exists, the addition of RO to a leaching system for water recycling may not be economically justified. The decrease in the wastewater treatment cost due to reduced flow is in many cases exceeded by the increased cost assessed on the basis of the strength of the waste stream. For treatment by evaporation ponds, RO offers a clear advantage over direct leachate disposal, reducing land area requirements fivefold or more and costs by 50% to 70%. Land application of both leachate and concentrate can take advantage of nutrients extracted from the biomass, and could be used in the production of additional biomass with some interesting possibilities for selective management of nutrients and fuel

246

B.M. Jenkins et al. / Fuel Processing Technology 81 (2003) 223246

composition. Large areas are required to provide equal loading compared to in-field leaching by rain washing. With adequate good quality irrigation water dilution, application area can be reduced to about one-tenth the original source area of the biomass and maintain salinity within acceptable limits.

References
[1] B.M. Jenkins, R.R. Bakker, J.B. Wei, Biomass and Bioenergy 10 (4) (1996) 177 200. [2] B.M. Jenkins, R.R. Bakker, R.B. Williams, L.L. Baxter, S.Q. Turn, P. Thy, M. Sime, C. Lesher, G. Sclippa, C. Kinoshita, Measurements of the fouling and slagging characteristics of banagrass (Pennisetum purpureum) following aqueous extraction of inorganic constituents, in: R.P. Overend, E. Chornet (Eds.), Making a Business from Biomass in Energy, Environment, Chemicals, Fibers and Materials, Pergamon, Oxford, UK, 1997, pp. 705 718. [3] D.C. Dayton, B.M. Jenkins, S.Q. Turn, R.R. Bakker, R.B. Williams, D. Belle-Oudry, L.M. Hill, Energy and Fuels 13 (4) (1999) 860 870. [4] B.M. Jenkins, R.B. Williams, R.R. Bakker, S. Blunk, D.E. Yomogida, W. Carlson, J. Duffy, R. Bates, K. Stucki, V. Tiangco, WayZ, Combustion of leached rice straw for power generation, in: R.P. Overend, E. Chornet (Eds.), Biomass: A Growth Opportunity in Green Energy and Value-Added Products, Pergamon, Oxford, 1999, pp. 1357 1363. [5] S.Q. Turn, C.M. Kinoshita, D.M. Ishimura, Biomass and Bioenergy 12 (1997) 241. [6] Z. Amjad, Reverse Osmosis: Membrane Technology, Water Chemistry, and Industrial Applications, Van Nostrand Reinhold, New York, 1993. [7] Anon, Product Specifications, Desal-3VLP, Desalination Systems, Escondido, CA, 1990. [8] C.N. Sawyer, P.L. McCarty, G.F. Parkin, Chemistry for Environmental Engineering, McGraw-Hill, New York, 1994. [9] T.M. ONeail, O.E. Kirchner, W.J. Day, Achieving high recovery from brackish water with seeded reverse osmosis systems, Proceedings 42nd International Water Conference, Engineers Society of Western Pennsylvania, Pittsburg, PA, 1981. [10] R.R. Bakker, Biomass fuel leaching for the control of fouling, slagging, and agglomeration in biomass power generation, unpublished PhD dissertation, University of California, Davis, 2000. [11] R.J. Ray, Cost estimates, reverse osmosis/part VI, in: W.S. Ho, K.K. Sirkar (Eds.), Membrane Handbook, Chapman & Hall, New York, 1992. [12] J.D. Mannapperuma, Residual management in fruit processing plants, in: L. Somogyi, H.S. Ramaswamy, T.H. Hui (Eds.), Processing of FruitsScience and Technology, Biology, Principles, and Applications, vol. 1. Technomic Publishing, Lancaster, PA, 1996. [13] R.S. Ayers, D.W. Westcot, Water quality for agriculture, Irrigation and Drainage Paper, vol. 29. Food and Agriculture Organization of the United Nations, Rome, 1985. [14] Anon, A management plan for agricultural subsurface drainage and related problems on the westside of the San Joaquin Valley, Final Report of the San Joaquin Valley Drainage Program, California Department of Water Resources, Sacramento, CA, 1990. [15] B.M. Jenkins, Atmospheric pollutant emission factors from open burning of agricultural and forest biomass by wind tunnel simulations, Final Report (3 Vols.), CARB Project A932-126, California Air Resources Board, Sacramento, California, 1996. [16] D.E. Yomogida, B.M. Jenkins, Nutrient replacement for rice fields under straw removal, soil incorporation, and burning management, in: R.P. Overend, E. Chornet (Eds.), Making a Business from Biomass in Energy, Environment, Chemicals, Fibers and Materials, Pergamon, Oxford, 1997, pp. 257 267.