Bo Ranneby

Methane emission from Swedish mires: National and regional budgets and dependence on mire vegetation

A national land use inventory program was used to estimate the flux of methane from Swedish mires. During the snow-free season in 1994, methane fluxes were monitored at 60% (n = 619) of the total number of permanent inventory plots for the four mire vegetation classes included. The methane flux was measured only once at each inventory plot using transparent static chambers, yielding 3157 (89%) accepted flux rates. The temperature during the measurement period was close to the long-term (30 year) average in the north and much warmer in the south. The precipitation was lower than the long-term average. Average flux rates, specific for each mire vegetation class and geographical region, varied between 8 and 238 mg CH4 m−2 d−1. The national averages ± SE (mg CH4 m−2 d−1) were hummock communities, 24±6; transitional fens, 15±2; short sedge fens, 49±4; and tall sedge fens, 97±22. To derive the annual flux, the number of days with a daily average air temperature over 5°C was used. The annual methane fluxes for each mire vegetation class and region varied between 1.5 and 40 g CH4 m−2 yr−1. The annual methane fluxes (average ± SE in g m−2 yr−1) for each mire vegetation class for the whole of Sweden were hummock sites, 4.9±1.3; transitional fens, 2.5±0.5; low sedge fens, 8.2±0.8; and tall sedge fens, 16.5±3.6. The methane flux (±95% confidence limits) in 1994 from the monitored mire types in Sweden was 0.29±0.05 Tg CH4. The low and tall sedge mires accounted for 96% of the methane emitted.

Temporal change estimation of mercury concentrations in northern pike (Esox lucius L.) in Swedish lakes

Adequate temporal trend analysis of mercury (Hg) in freshwater ecosystems is critical to evaluate if actions from the human society have affected Hg concentrations ([Hg]) in fresh water biota. This study examined temporal change in [Hg] in Northern pike (Esox lucius L.) in Swedish freshwater lakes between 1994 and 2006. To achieve this were lake-specific, multiple-linear-regression models used to estimate pike [Hg], including indicator variables representing time and fish weight and their interactions. This approach permitted estimation of the direction and magnitude of temporal changes in 25 lakes selected from the Swedish national database on Hg in freshwater biota. A significant increase was found in 36% of the studied lakes with an average increase in pike [Hg] of 3.7±6.7% per year that was found to be positively correlated with total organic carbon. For lakes with a significant temporal change the dataset was based on a mean of 30 fish, while for lakes with no temporal change it was based on a mean of 13 fish.

On the determination of time intervals between remeasurements of permanent plots

Abstract Remeasurement frequency in permanent sample plots is discussed within the framework of sampling with partial replacement (SPR) and based on data from the Swedish National Forest Inventory (NFI). Previously, Swedish NFI plots were measured every 5 years. The results indicate that it is possible to extend the time interval until the next remeasurement. An extension from 5 to 10 years increases variance by 5–13%, depending on the temporal correlation. By treating the clear-cut plots as a separate stratum with known area, there is a considerable gain in precision. With such a stratum and a measurement interval of 10 years, the variance decreases by 15–20%. By using an approach with intuitive estimators, and viewing the SPR estimator as a linear combination of these, it is possible to obtain additive tables with only minor loss in efficiency.

A general autoregressive model with Markov switching: Estimation and consistency

In this paper, a general autoregressive model with Markov switching is considered, where the autoregression may be of an infinite order. The consistency of the maximum likelihood estimators for this model is obtained under regularity assumptions. Examples of finite and infinite order autoregressive models with Markov switching are discussed. Simulation studies with these examples illustrate the consistency and asymptotic normality of the estimators.

Forecasting using locally stationary wavelet processes

Locally stationary wavelet (LSW) processes, built on non-decimated wavelets, can be used to analyse and forecast non-stationary time series. They have been proved useful in the analysis of financial data. In this paper, we first carry out a sensitivity analysis, then propose some practical guidelines for choosing the wavelet bases for these processes. The existing forecasting algorithm is found to be vulnerable to outliers, and a new algorithm is proposed to overcome the weakness. The new algorithm is shown to be stable and outperforms the existing algorithm when applied to real financial data. The volatility forecasting ability of LSW modelling based on our new algorithm is then discussed and shown to be competitive with traditional GARCH models.

Estimation of WTP with point and self-selected interval responses

In contingent valuation studies regarding willingness to pay (WTP) the respondents usually give an exact value as his/her WTP-value. Unfortunately, the non-response rate has a tendency to be quite high. As an attempt to reduce that rate the respondents will have a possibility to give a self-selected interval instead of a fixed value as their WTP. In this paper we will study different approaches to estimate the mean willingness to pay under these conditions. First we consider the nonparametric and a parametric approach where the intervals are treated as if the respondent gives an exact value but we cannot observe it. Next we will give a different interpretation of the intervals: Included in the respondent’s answer is information about his/her uncertainty about what would be a reasonable value of WTP. For illustration purposes we will use data from a small study in Bollnas municipality. In all three situations we first estimate the mean WTP and its standard error for those giving a positive answer and finally we add zero-responses.

On strong Hellinger consistency of posterior distributions

We establish a sufficient condition ensuring strong Hellinger consistency of posterior distributions. We also prove a strong Hellinger consistency theorem for the pseudoposterior distributions based on the likelihood ratio with power 0<α<1, which are introduced by Walker and Hjort [2001 ‘On Bayesian Consistency’, J. R. Statist. Soc., B 63, 811–821]. Our result is an extension of their theorem for α=½.

Nonparametric and Probabilistic Classification Using NN-balls with Environmental and Remote Sensing Applications

National and international policies today require environmental follow-up systems that detect, in a quality assured way, changes over time in land use and landscape indicators. Questions related to environmental health and spatial patterns call for new statistical tools. We present in this paper some new developments on the classification of land use by using multispectral and multitemporal satellite images, based on techniques of nearest neighbour balls. The probabilistic classifiers introduced are useful for measuring uncertainty at pixel level and obtaining reliable area estimates locally. Also some theoretical considerations for the reference sample plot method (today named k-NN method in natural resource applications) are presented.

A diameter growth index method for standardization of forest data inventoried at different dates

A model of tree diameter growth during the growth season was developed to standardize the forest data from different inventory dates. A diameter growth index is defined and its relation with the number of growing days is established. Parameters characterizing this relationship are estimated using accurate tree ring width measurements and other stand and tree related variables from the Swedish National Forest Inventory data for 1989. The residual analyses show that the model is an effective expression of the actual growth and the results are consistent with an earlier used tree growth model. Some climatic data are tested for model improvement, but no essential improvement is gained given the data available.