Search

We present OH observations made in Amazonas, Brazil during the Green Ocean Amazon campaign (GoAmazon2014/5) from February to March of 2014. The average diurnal variation of OH peaked with a midday (10:00–15:00) average of 1.0 × 106 (±0.6 × 106) molecules cm−3. This was substantially lower than previously reported in other tropical forest photochemical environments (2–5 × 106 molecules cm−3) while the simulated OH reactivity was lower. The observational data set was used to constrain a box model to examine how well current photochemical reaction mechanisms can simulate observed OH. We used one near-explicit mechanism (MCM v3.3.1) and four condensed mechanisms (i.e., RACM2, MOZART-T1, CB05, CB6r2) to simulate OH. A total of 14 days of analysis shows that all five chemical mechanisms were able to explain the measured OH within instrumental uncertainty of 40% during the campaign in the Amazonian rainforest environment. Future studies are required using more reliable NOx and VOC measurements to further investigate discrepancies in our understanding of the radical chemistry in the tropical rainforest. © 2021. American Geophysical Union. All Rights Reserved.

In the model of fault tolerant decision trees introduced by Kenyon and Yao, there is a known upper bound E on the total number of queries that may be faulty (i.e., get the wrong bit). We consider this computational problem: Given as input the truth table of a function f: {0,1}ⁿ → {0,1} and a value of E, find the minimum possible height (worst-case number of queries) of any decision tree that computes f while tolerating up to E many faults. We design an algorithm for this problem that runs in time Õ(binom(n+E,E)⋅(2E+3)ⁿ), which is polynomial in the size of the truth table when E is a constant. This generalizes a standard algorithm for the non-fault tolerant setting.

This study aimed to investigate the differences in external match load between professional and semi-professional footballers, and also aimed to investigate whether periods of fixture congestion throughout the season had an effect on the external match load of players at either the professional or semi-professional level. This study consisted of data from 51 football players, 21 professional and 30 semi-professional footballers, playing in the 2019/2020 football season. The data collected was obtained via MEMS (microelectromechanical systems) devices, which measured the players’ total distance, high-speed distance, accelerations, decelerations and player load. Once the external match load data was quantified, a comparison between playing levels took place using a univariate ANOVA. A two-way repeated measures ANOVA was used to examine if significant differences existed in external match load variables across player performance level (2 levels) and time of the season (3 levels) during periods of time when teams experienced fixture congestion. This study found that professional players travelled significantly greater distances in a 90 minute match (10.93 ± 2.46 vs 9.02 ± 1.56 km respectively; P<0.001). No differences in high-speed distance were observed between playing level (P=0.70), whereas semiprofessional players recorded significantly greater player load value than the professional players (88.6 ± 12.2 vs 68.8 ± 18.9% respectively; P<0.001). Periods of fixture congestion were not found to significantly affect any of the match load variables at either playing level despite the time of the season. In conclusion, neither playing level was found to exhibit a superior level of external match load. The other major finding of this thesis was that fixture congestion did not affect match load. Further research is required to quantify and compare the external match load at the non-elite professional and semi-professional level of football, as these levels of football are la