Author-produced version of the article published in Waste Management Original publication available at www.elsevier. com -doi:10.1016/j.wasman.2009.10.018 2 Abstract 1 This paper aimed to compare household waste, separated pig solids, food waste, pig slaughterhouse sludge 2 and green algae regarding processes ruling nitrogen dynamic during composting. For each waste, three 3 composting simulations were performed in parallel in three similar reactors (300L), each one under a constant 4 aeration rate. The aeration flows applied were comprised between 100 and 1100 L/h. The initial waste and the 5 compost were characterized through the measurements of their contents in dry matter, total carbon, Kjeldahl and 6 total ammoniacal nitrogen, nitrite and nitrate. Kjeldahl and total ammoniacal nitrogen and nitrite and nitrate 7were measured in leachates and in condensates too. Ammonia and nitrous oxide emissions were monitored in 8 continue. The cumulated emissions in ammonia and in nitrous oxide were given for each waste and at each 9 aeration rate. The paper focused on process of ammonification and on transformations and transfer of total 10 ammoniacal nitrogen. The parameters of nitrous oxide emissions were not investigated. The removal rate of 11 total Kjeldahl nitrogen was shown being closely tied to the ammonification rate. Ammonification was modelled 12 thanks to the calculation of the ratio of biodegradable carbon to organic nitrogen content of the biodegradable 13 fraction. The wastes were shown to differ significantly regarding their ammonification ability. Nitrogen 14 balances were calculated by subtracting nitrogen losses from nitrogen removed from material. Defaults in 15 nitrogen balances were assumed to correspond to conversion of nitrate even nitrite into molecular nitrogen and 16 then to the previous conversion by nitrification of total ammoniacal nitrogen. The pool of total ammoniacal 17 nitrogen, i.e. total ammoniacal nitrogen initially contained in waste plus total ammoniacal nitrogen released by 18 ammonification, was calculated for each experiment. Then, this pool was used as the referring amount in the 19 calculation of the rates of accumulation, stripping and nitrification of total ammoniacal nitrogen. Separated pig 20 solids were characterised by a high ability to accumulate total ammoniacal nitrogen. Whatever the waste, the 21 striping rate depended mostly on the aeration rate and on the pool concentration in biofilm. The nitrification rate 22 was observed as all the higher as the concentration in total ammoniacal nitrogen in the initial waste was low. 23Thus, household waste and green algae exhibited the highest nitrification rates. This result could mean that in 24 case of low concentrations in total ammoniacal nitrogen, a nitrifying biomass was already developed and that 25 this biomass consumed it. In contrast, in case of high concentrations, this could traduce some difficulties for 26 nitrifying microorganisms to develop. 27

This paper aims to compare household waste, separated pig solids, food waste, pig slaughterhouse sludge 2 and green algae regarding their biodegradability, their stabilization kinetics and their temperature rise during 3 composting. Three experiments in lab-scale pilots (300L) were performed for each waste, each one under a 4 constant aeration rate. The aeration rates applied were comprised between 100 and 1100 L/h. The 5 biodegradability of waste was expressed as function of dry matter, organic matter, total carbon and chemical 6 oxygen demand removed, on one hand, and of total oxygen consumption and carbon dioxide production on the 7 other. These different variables were found closely correlated. Time required for stabilization of each waste was 8 determined too. A method to calculate the duration of stabilization in case of limiting oxygen supply was 9proposed. Carbon and chemical oxygen demand mass balances were established and gaseous emissions as 10 carbon dioxide and methane were given. Finally, the temperature rise was shown to be proportional to the total 11 mass of material biodegraded during composting. 12 13

Environmental assessment of Municipal Solid Waste (MSW) management is essential. Life Cycle Assessment (LCA) is a powerful and widely used method, which implements causal chains (impact pathways) between the studied processes and their environmental impacts. However, in waste management, the method presents some weaknesses. For example, there is no impact category related to odour, whose assessment is nevertheless essential, especially when the organic fraction of waste is concerned. Odour interferes with human welfare and comfort. Sometimes, it can become a nuisance and be described as a "socio-environmental" impact. To integrate the impact of odour in waste management plans, it is necessary to build an odourimpact pathway. The aim of this paper is to present a first attempt to build such an impact pathway up to the so-called midpoint step (i.e. the level of discomfort to human beings). The methodology we developed is based on the cause/effect chain according to the descriptors of the Site Dependent approach. Unlike classical LCA, the classification step is more important and characterization is aimed at computing the characterization factor. The change in this classification step allows for working on the occurrence of odour impacts. To determine impact occurrence, it is necessary to integrate local conditions into odour assessment. This was done using (i) the USEtox model in which local conditions to assess odour impacts are integrated and (ii) the framework of a new methodology that takes into account background concentrations). The methodology was implemented in a case study, i.e. by computing atmospheric emission of ethyl benzene during composting (2.93.10-2 kg.d-1). The characterization factor for ethyl benzene was equal to 3.02.10-3 kg eq. Benzene per /kg emitted ethyl benzene. The daily emission of ethyl benzene generated an odour impact equal to 6.6.10-5 kg eq. benzene. With that first odour mid-point impact, we paved the way for the construction of a whole odour pathway (going up to end-point impacts or damages). However, several limits were identified such as data availability, the model under use and the use of average daily data which is less relevant than emission peaks. We should also recall that our methodology is not intended for predicting nuisance likely to disturb populations living nearby the facility. Its first objective is to provide an indicator that fits with LCA methodology in order to help local decision-makers to differentiate waste management scenarios based on exhaustive LCA.

No abstract