### Table 5. Overview of the meteorological conditions over Paris and maximum surface ozone concentrations in representative locations during the IOP 2. The background and plume concentrations are from the AIRPARIF stations in Montge en Goele and Rambouil- let, respectively. The times were the maximum of ozone occured are given in parentheses.

2000

"... In PAGE 8: ... 4 IOP2 of ESQUIF: Data analysis The ESQUIF experiments performed during Summer 1998 have been designed to address more specifically the follow- ing questions: (i) what are the relative contributions of local production and long-range transport to the ozone concentra- tions measured over the Paris area during pollution events and (ii) what is the impact of the vertical boundary layer structure on ozone build-up over the Paris area? In this section, we per- form preliminary analyses in order to illustrate how ESQUIF data can be used to address these questions. In this paper, we focus on IOP 2 which lasted three days, from Friday to Sunday 7-9 August 1998 (hereafter called 7A98, 8A98 and 9A98, respectively, Table5 ). This period (including 10 and 11 August 1998) corresponded to the high- est ozone concentrations recorded during the whole summer of 1998.... ..."

Cited by 1

### Table 3. Conections betwen IOPE and proceses

2006

"... In PAGE 35: ...DF. The former case is for other languages such as KIF3 and PDL. This type of representation of the IOPE parameters needs to be atached to the proces. This is done with the properties shown in Table3 . The links from a proces to its parameters implicitly gives them scope.... ..."

### Table 4: FTP and TVTP Models for IOP

"... In PAGE 11: ... Thus, these latter models appear to explain the univariate dynamics of IOP. The sign and significance of the mean growth rates for the Hamilton FTP model in Table4 indicate classical business cycle behaviour. The data can be classified into positive and negative growth rate regimes, with a mean growth rate of .... In PAGE 11: ... Note that the model selection criteria AIC and SIC cannot be compared with the linear AR specification as an indication of regime switching non-linearity because of the non-standard testing conditions that are involved (see Hamilton and Perez- Quiros, 1996). The TVTP results for the single indicator and two cases with two indicators are shown in Table4 . The estimates of the regime dependent means associated with the TVTP logistic estimations are statistically significant and again indicate classical business cycle behaviour in IOP.... In PAGE 12: ... Two-indicator TVTP models were also considered. Both such models shown in Table4 use IR as the leading indicator during contractions (regime 0), since in terms both of the magnitude and significance of its estimated slope coefficient b11 this variable apparently provides most leading indicator information in this regime. During expansions (regime 1), the difference of TS and also housing starts (HS) are considered, the former having the largest and most significant slope coefficient in that regime with the latter also having a highly significant coefficient.... In PAGE 13: ... The FTP model itself provides very little evidence of this last recession. Essentially, the reason appears to be that the two-regime Markov-switching models of Table4 associate recession with very strong quarterly declines in industrial production. Although the 1990s recession evidenced a sustained period of production declines, it did not contain any single quarter with a decline of the magnitude seen in each of the three previous IOP recessions (mid-1970, early 1980s and mid-1980s).... ..."

### Table 4: FTP and TVTP Models for IOP

"... In PAGE 11: ... Note that the model selection criteria AIC and SIC cannot be compared with the linear AR specification as an indication of regime switching non-linearity because of the non-standard testing conditions that are involved (see Hamilton and Perez- Quiros, 1996). The TVTP results for the single indicator and two cases with two indicators are also shown in Table4 . The estimates of the regime dependent intercepts associated with the TVTP logistic estimations are statistically significant and again indicate classical business cycle behaviour in IOP.... In PAGE 12: ... Two-indicator TVTP models were also considered. Both such models shown in Table4 use IR as the leading indicator during contractions (regime 0), since in terms both of the magnitude and significance of its estimated slope coefficient b0 1 this variable apparently provides most leading indicator information in this regime. During expansions (regime 1), the difference of TS and also housing starts (HS) are considered, the former having the largest and most significant slope coefficient in that regime with the latter also having a highly significant coefficient.... In PAGE 13: ... (1997). Hence, the shading provides some benchmark for the performance of the models of Table4 in terms of capturing historical recessions. A few points about Figure 2 are worth noting.... In PAGE 13: ... The FTP model itself provides very little evidence of this last recession. Essentially, the reason appears to be that the two-regime Markov-switching models of Table4 associate recession with very strong quarterly declines in industrial production. Although the 1990s recession evidenced a sustained period of production declines, it did not contain any single quarter with a decline of the magnitude seen in each of the three previous IOP recessions (mid-1970, early 1980s and mid-1980s).... ..."

### Table 2: Linear Models for IOP with one Leading Indicator

"... In PAGE 9: ... No lag specification is included in Table 1 for a TVTP specification using the level of CBIO because no satisfactory model could be found. Table 1: Leading Indicator Lags for IOP Variable Linear TVTP K * K 1 * K 0 * M0 5 2 2 IR 5 3 4 SP 3 3 3 HS 5 4 7 DY 3 3 1 TBY 5, 7 8 6 LR 3 3 3 Level TS 4 6 4 Difference TS 7 5 6 Level CBIO 1 n/a n/a Difference CBIO 3 3 7 Estimation and Diagnostic Test Results First of all we briefly examine the linear model estimation and diagnostic test results presented in Table2 . The LM residual diagnostics for the linear univariate AR(1) specification do not show strong evidence of model ... ..."

### Table 1. Results for 3 test matrices using multiple right hand side vectors. The table reports average number of CG iterations for 50 runs, using di erent values of the memory parameter m and di erent storage schemes. The initial point for every CG iteration is x(0) = 0. A10 A11 A20 m IOP=1 IOP=2 IOP=1 IOP=2 IOP=1 IOP=2

"... In PAGE 7: ...12) we proceed as follows: i) the quasi-Newton preconditioner is computed with the information gathered by the CG method while solving the rst system Ax = b1; ii) the remaining 50 problems are solved using this preconditioner. We report in Table1 the average number of CG iterations (rounded to the nearest integer) required to solve the 50 preconditioned systems, for various values of m. Both strategies for forming the preconditioner were used: uniform sampling (IOP = 1) and the last correction pairs (IOP = 2).... ..."