The data analysis will be done with various computer-based methods. Some analysis scripts are already available and only need minor adjustments to be applicable to CHAOS flight data. It is planned to use the acquired data for \change[4]{master's and} bachelor's theses.
\change[3]{The acquired flight data can be used to determine the position of the Regener-Pfotzer maximum. For this, only the particle measurements of the \ac{BGO} scintillation detector are needed to calculate a total particle count rate. The height can be calculated using the pressure measurements.}
Different particle species can be identified by using the measured energy losses of the \acfp{SSD} and scintillation detector. \replaced{Data analysis methods such as the so-called "fish plots" and the $\frac{\mathrm{d}E}{\mathrm{d}x}$-$E$-method which are established at the Department for Extraterrestrial Physics will be used. Especially the $\frac{\mathrm{d}E}{\mathrm{d}x}$-$E$-method is often used by the scientific community in general.}{Established data analysis methods such as "fish plots" and the $\frac{\mathrm{d}E}{\mathrm{d}x}$-$E$-method. These methods are widely used at the Department of Extraterrestrial Physics and the scientific community in general.} The further separation of light and heavy particles using the Cherenkov detector as threshold detector will be tested for the first time on real flight data.
\change[3]{Most of the data analysis will take place after the launch campaign in Kiruna because an in-depth analysis of the data is needed. It is very likely that some data cleaning will be needed. The calculation of a count rate can be done in Kiruna.}
\change[4]{The needed evaluation of the calibration measurements performed before the balloon flight will also be done as part of the data analysis after the launch campaign.}
This section \replaced[5]{includes}{will include} a description of the flight preparation, flight performance, recovery of the experiment and post flight activities.
\change[5]{The first six team members arrived at \ac{Esrange} Space Center in the evening of the 27th September 2024.
On 28th September, the instrument was unpacked and a first functionality test was performed with a direct connection from the ground station to the \ac{CHAOS} instrument. All systems were nominal. In the afternoon a first individual experiment test was performed on the gondola using first the hardline ethernet connection and then the \ac{RF} connection. \ac{CHAOS} could send and receive data throughout the whole test and reconnected itself after simulated connection losses for both hardline and \ac{RF}. Nevertheless, \ac{SSC} could not see the data transfer in their systems. After testing both the hardline and \ac{RF} connection for a second time, the test was terminated. It turned out that the problem was on \ac{SSC}'s side. The following night was used for a longer ground measurement at \ac{Esrange}.
On 29th September, a second individual experiment test was performed on the gondola after \ac{SSC} fixed their problems. The connection via \ac{RF} worked and \ac{CHAOS} could handle connection dropouts caused by disconnecting cables on the gondola and turining off \ac{SSC}'s ethernet software. After turning back on the software, a big spike in the data rate of the downlink could be observed. This was first attributed to \ac{CHAOS} by \ac{SSC} but it turned out that \ac{SSC}'s own buffer caused the spike. The hardline connection was not tested again. After this test the grounding strip of the instrument was disconnected. \ac{SSC} did not want the minus pole of the battery to be connected to the chassis. Apparently, there was some miscommunication between \ac{ZARM} and \ac{SSC} before the launch campaign. that is why the \ac{ZARM} had approved the instrument's design before.
On 30th September, the final integration onto the gondola took place. M8 screws with a torque of 7.5\,Nm were used. Furthermore, the instrument was covered in rescue foil for additional thermal control. An interference test with all experiments on the gondola was performed. First, all experiments were turned on, tested and turned off individually using the \ac{RF} connection. Afterwards all instruments were tested at the same time. \ac{CHAOS} behaved nominally.
On 1st October 2024, the \ac{FCT} was performed. The gondola was moved outside the assembly hall by the Hercules vehicle and all instruments were tested using the internal battery and the \ac{RF} connection. For \ac{CHAOS} all systems were nominal. The gondola of \ac{BEXUS} mission 35 was ready for launch and can be seen in figure \ref{fig:data_analysis:gondola_35}.}
\change[5]{On 2nd October 2024, the stratospheric balloon mission \ac{BEXUS} 35 was launched from \ac{Esrange} Space Center. The countdown had two holds due to unfavorable wind conditions. At 9:53\,am local time the balloon finally launched. Its flight trajectory can be seen in figure \ref{fig:data_analysis:trajectory}. The balloon reached its floating altitude of around 26 to 27\,km at 11:23\,am after 1.5~hours. It stayed at this altitude for more than 3.5~hours and was cut off at 3:03\,pm. Due to a flight path over inhabited areas which did not allow for the balloon to be cut off, the floating time was unusually long. \ac{SSC} originally only guaranteed a floating time of one hour. This means that the floating time was more than three times as long as expected leading to a lot of gathered statistics. After half an hour of descent, the balloon landed in Finland with the loss of radio connection at 3:31\,pm.}
\change[5]{The most important housekeeping data is shown in figures \ref{fig:data_analysis:hk_pressure}, \ref{fig:data_analysis:hk_temp} and \ref{fig:data_analysis:hk_counts}. The first altitudes shown in figure \ref{fig:data_analysis:hk_pressure} are calculated from \ac{CHAOS}'s external pressure sensor using the U.S. Standard Atmosphere \citep{atmosphere1976us}. In comparison, the \acf{GNSS} altitudes provided by \ac{SSC} are shown as well. The difference in the altitudes can be attributed to a systematic error of \ac{CHAOS}'s external pressure sensor.}
\change[5]{The data was stored in binary format on a SD card within the instrument. Furthermore, the binary data was downlinked to the ground station. On ground, the binary data was transformed into an \acs{ASCII} data file. The binary file had a size of around 257\,MB and the \acs{ASCII} file of 790\,MB.\\
Prior to the launch campaign, it was suspected that the instrument would be sensitive to strong shocks and vibrations occurring during flight. Figure \ref{fig:data_analysis:hk_counts} shows the count rates of representative detector channels. Spikes in the counts can be observed for the launch and cutoff of the balloon. These can be handled with ease making no elaborate data filtering techniques necessary.}
\change[5]{The \ac{CHAOS} instrument showed no unusual behaviour during the whole flight. All systems were nominal and worked reliably. There were some minor problems with the ethernet connection to the instrument during ascent. \ac{SSC} apparently had some pointing issues with their radar dishes. There was a longer connection loss of around 30~minutes at 12:30\,pm. This is visible as data gaps in figure \ref{fig:data_analysis:hk_counts}. The loss was caused by power problems of the antennas. Considering the unusually long floating time, these connection losses are no problem for the data evaluation. The \ac{CHAOS} instrument could gather enough statistics. Furthermore, all data was saved to an onboard SD card, which was successfully readout in Kiel. This means no data for post flight analysis was lost due to the disconnects.}
\change[5]{Figure \ref{fig:data_analysis:hk_temp} shows the temperatures measured by the internal and external temperature sensor during flight. The temperature sensors T8\_R1 and T12\_R1 are placed inside the pressure housing on the base plate and the detector bracket, respectively. Tbgo is placed on the \ac{BGO} crystal. It can be seen that the temperatures reached a plateau while the instrument was on ground and during the floating phase. There is a small decrease in temperature during the floating phase which can be attributed to a small decrease in altitude. In the stratosphere lower altitudes correspond to colder temperatures. During the floating phase, all temperatures within the instrument stayed above 10\,$^\circ$C. The \ac{HV} remained stable throughout the whole flight with no drop in voltage due to cold temperatures which would have impacted the performance of the Cherenkov detector B negatively.}
\subfigure[BX35 gondola during recovery.]{\includegraphics[width=0.49\linewidth]{images/data_analysis/recovery3.jpg}\label{fig:data_analysis:recovery3}}
\subfigure[BX35 gondola during recovery.]{\includegraphics[width=0.49\linewidth]{images/data_analysis/recovery2.jpg}
\label{fig:data_analysis:recovery2}}
\subfigure[\ac{CHAOS} on the gondola during recovery.]{\includegraphics[width=0.49\linewidth]{images/data_analysis/recovery1.jpg}
\change[5]{The gondola of \ac{BEXUS} mission 35 was recovered by an helicopter on 3rd October 2024 and showed no major damage. Apparently, the landing must have been very soft. The \ac{CHAOS} instrument itself also showed no visible damage from the outside. Figure \ref{fig:data_analysis:recovery} shows both the gondola and the instrument during recovery. After recovery, the gondola was transported back to \ac{Esrange} by truck were it arrived on 4th October 2024.}
\change[5]{After its return to \ac{Esrange}, the instrument was inspected for visible damage in more detail. The inside and outside of the insulation box showed no visible damage. Furthermore, the mechanical interface showed no strong visible deformations. To check for its functionality, the instrument was turned on for a ground measurement in the assembly hall. At first sight, the instrument showed the same behaviour as for ground measurements prior to the balloon flight. After issuing the command to power the \acf{HV}, the counts of the Cherenkov detector rose showing the expected behaviour but the housekeeping value for the \ac{HV} was still 0\,V. Therefore, the \ac{HV} status LED did not work properly. The fact that the Cherenkov detector worked means that there must be a problem with the \ac{HV} monitoring and not its generation. During flight, no failure in the \ac{HV} monitoring was observed. This most likely means that it broke during or after landing. It was decided to further investigate the \ac{HV} monitoring back in Kiel when the pressure housing is opened. The instrument was packed up for transportation back to \ac{ZARM} in Bremen where it arrived on 23rd October 2024.\\From Bremen \ac{CHAOS} was then sent to Kiel. It arrived on 12th November 2024. Figure \ref{fig:data_analysis:post_kiel} shows pictures from the inspection performed after the instrument's arrival. No damage of the insulation box and pressure housing from the transportation to Kiel could be detected. The pressure housing was opened and the sensor head underwent a visual inspection. First, a protective cap was placed on \ac{SSD} A. The \ac{HV} monitoring cable was disconnected which explains the \ac{HV} housekeeping value of 0\,V in the post-flight measurement. A number of SSMC connectors were also only loosely connected. Furthermore, a nut which fastened the lid of the \ac{E-Box} was loose. The cause are most likely vibrations. Therefore, for future missions, all critical screws and connections should be fixated with glue. After the inspection, the instrument was powered with 16\,V and the SD card was read out using a \ac{USB} connection.}
This section \replaced[5]{includes}{will include} the technical results and scientific data evaluation as well as an outlook for further activities. The scientific success in respect to the objectives will be estimated.
\caption{Regener-Pfotzer Maximum for \ac{CHAOS}'s stratospheric balloon flight in ACD-coincidence for all events (blue), events where the Cherenkov detector B was triggered (green) and events where it was not triggered (orange).}
\change[5]{Objective 1.1 states that \ac{CHAOS} shall measure primary \acfp{GCR} above the so-called Regener-Pfotzer Maximum (RP-Maximum) and separate the light and heavy particles within them using the Cherenkov detector. The primary \acp{GCR} reaching the Earth's atmosphere from outer space cause a shower of secondary ionizing particles. The intensity of this shower depends on the altitude within the atmosphere. For the measurement of primary \acp{GCR}\ac{CHAOS} ideally reaches altitudes above the RP-Maximum. Figure \ref{fig:data_analysis:rp_max} shows the count rate of ionizing particle radiation during the ascent of the balloon in bins of four minutes. The error bars of the counts are the standard deviation based on Poisson statistics. An ACD-coincidence is used, meaning that a particle must have triggered the \acfp{SSD} A and C as well as the \ac{BGO} scintillation detector D. The \acp{SSD} are placed above and below the Cherenkov detector B. With this coincidence the vertical radiation intensity is considered. The fits for the determination of the Regener-Pfotzer Maximum are based on \acf{GP} Regression which is a machine learning based approach to data fitting. For further information refer to \citet{williams2006gaussian} and \citet{scikit-learn}. The solid lines in figure \ref{fig:data_analysis:rp_max} are the predictions from the \ac{GP} Regression. As altitude of the RP-Maximum the position of the highest predicted count rate is taken. The lighter shaded areas mark the 95\,\% confidence interval of the predictions. From the range of these areas the error bars of the given RP-Maxima are derived. It has to be kept in mind that these are no statistical errors. They can be interpreted as measure of how pronounced the corresponding RP-Maximum is. The blue data shows all events in ACD-coincidence with the RP-Maximum at approximately 17\,km. \ac{CHAOS} reached altitudes above this maximum. The first part of objective 1.1 is therefore met. To the data in ACD-coincidence the Cherenkov Detector B can be applied. For events where the Cherenkov detector was triggered (green) the RP-Maximum shifts to a lower altitude. If only events are considered where the Cherenkov detector was not triggered (orange), the maximum shifts to a higher altitude of almost 20\,km. The maximum is also less pronounced. Furthermore, it can be seen that there are always more particles which trigger the Cherenkov detector than particles which do not. The results show that the exact position of the RP-Maximum heavily depends on the energy of the considered particles because they need a certain energy to exceed the Cherenkov detector's threshold velocity.}
\caption{Solar modulation of the Regener-Pfotzer Maximum. The International Sunspot Number $S_\text{n}$ is given as monthly mean and as averaged value taking the current month as well as the six previous and six following months into account \citep{sidc}. It is a proxy value for the solar activity. The further altitudes of the RP-Maximum next to \ac{CHAOS}'s value are taken from \citet{ambrovzova2023latitudinal} and previous \ac{BEXUS} missions from Kiel University \citep{moeller2013phd,doensdorf2014phd}.}
\change[5]{\ac{CHAOS}'s stratospheric balloon flight offers the chance to compare the altitude of the RP-Maximum for different solar activities. The Sun's activity follows an 11-year cycle \citep{hathaway2015solar}. The \ac{BEXUS} 35 balloon flight took place in October 2024, which is close to a solar maximum. \citet{ambrovzova2023latitudinal} provides results for the RP-Maximum's altitude in September 2019, which was during a solar minimum and the PhD theses of \citet{moeller2013phd} and \citet{doensdorf2014phd} from Kiel University give values for times of intermediate solar activity. All these altitudes of the RP-Maximum are based on stratospheric balloon flights from Kiruna. The results from \citet{moeller2013phd} and \citet{doensdorf2014phd} are even based on previous \ac{BEXUS} missions. Differences in the altitude caused by different geomagnetic shielding of primary \acfp{GCR} at different geographic positions can be neglected because all flights started at the same place. Figure \ref{fig:data_analysis:rp_max_solar_modulation} shows the altitude of the RP-Maximum as well as the International Sunspot Number $S_\text{n}$ in the years 2011 to 2024 \citep{sidc}. The number of sunspots is used as a measure of the Sun's magnetic activity \citep{hathaway2015solar}. An anticorrelation of the RP-Maximum's altitude to the sunspot number and therefore solar activity can be observed. This behavior is expected. The primary \acp{GCR} reaching the top of the Earth's atmosphere show a reduced particle flux for energies below around 10\,GeV/nuc in times of high solar activity \citep{abe2016measurements}. This is caused by the stronger magnetic field of the sun, which is carried outwards by the solar wind. More of the low-energy primary \acp{GCR} are prevented from entering the solar system \citep{hathaway2015solar}. The primary \acp{GCR} reaching the Earth's atmosphere cause a shower of secondary particles. The altitude of the shower's maximum depends on the energy of the particle starting the cascade. For primary particles with lower energies, the shower maximum is reached in higher altitudes \citep{letessier2011ultrahigh}. The position of the RP-Maximum is determined by the superposition of all showers occurring at the same time. With less showers caused by low-energy particles, the RP-Maximum can be found at lower altitudes.\\
It has to be emphasized that the discussion of the solar modulation in this section is only of qualitative nature. The determined altitude of the RP-Maximum depends on the used detector's response to different particle types and their energies.}
\caption{Simulation of the energy losses of particles entering the sensor head from above. The x-axis shows the energy loss in the \ac{BGO} and the y-axis the minimum energy loss of \ac{SSD} A and \ac{SSD} C divided by the maximum energy loss of \acp{SSD} E1, E2 and E3. The two-dimensional histogram shows the sum of counts for protons, helium and electrons. The contour lines show the regions where each particle species appears in this plot.}
\caption{Simulation of the energy losses of particles entering the sensor head from above. The x-axis shows the energy loss in the \ac{BGO} and the y-axis the energy loss in \ac{SSD} C. The two-dimensional histogram shows the sum of counts for protons, helium and electrons. The contour lines show the regions where each particle species appears in this plot.}
\change[5]{A Geant4 \citep{AGOSTINELLI2003250} simulation was used to better identify the particle types and energies in the measurements. For this simulation the sensor head and the pressure housing around it were modeled. In the next step, particles of different types and energies as well as their interactions with the detector were simulated. Visual representations of the simulation results are shown in figures \ref{FishSim} and \ref{CDSim}. To understand these plots, one needs to know that heavy particles loose energy in accordance to their squared charge number and antiproportional to their kinetic energy as described by the Bethe-Bloch formula \citep{Bethe}.
This means for example that a particle coming from above looses a small amount of energy $E_1$ in the \ac{SSD} C above the \ac{BGO}, a lot of energy in the \ac{BGO} and an energy $E_2$ in the \ac{SSD} E below the \ac{BGO}. $E_2$ is slightly bigger than $E_1$, resulting in a quotient $E_1/E_2<1$. As the primary kinetic energy of the particle decreases, the quotient of $E_1/E_2$ gets smaller. This leads to the tails that can be seen in figure \ref{FishSim} for protons and helium. Electrons however can not be described by the Bethe-Bloch formula as they are too light. Their different behavior can be seen in figures \ref{FishSim} and \ref{CDSim} where they overlap with the protons which is the reason for our use of the Cherenkov detector.
In figure \ref{CDSim}, the proton tail going from the upper left corner to the right are protons which stop in the \ac{BGO}. The more kinetic energy they have the less they lose in \ac{SSD} C. This continues until the particles penetrate the \ac{BGO} at which point the energy loss in the \ac{BGO} gets smaller as well, until both energy losses hit a minimum for so-called \acfp{MIP}.}
\subfigure[Events where the Cherenkov detector did not trigger.]{\includegraphics[trim = 5 5 5 30, clip,width=0.65\linewidth]{images/data_analysis/datsim-fish_Blower20.png}\label{fig:data_analysis:fish_Blower20}}
\subfigure[Events where the Cherenkov detector did trigger.]{\includegraphics[trim = 5 5 5 30, clip,width=0.65\linewidth]{images/data_analysis/datsim-fish_Bhigher20.png}\label{fig:data_analysis:fish_Bhigher20}}
\caption{Recreation of figure \ref{FishSim} with flight data under trigger conditions A, C, AeqC, and D. The simulation results are shown as contour lines.}
\subfigure[Events where the Cherenkov detector did not trigger.]{\includegraphics[trim = 5 5 5 30, clip,width=0.65\linewidth]{images/data_analysis/datsim-chaos_Blower20.png}\label{fig:data_analysis:chaos_Blower20}}
\subfigure[Events where the Cherenkov detector did trigger.]{\includegraphics[trim = 5 5 5 30, clip,width=0.65\linewidth]{images/data_analysis/datsim-chaos_Bhigher20.png}\label{fig:data_analysis:chaos_Bhigher20}}
\caption{Recreation of figure \ref{CDSim} with flight data under trigger conditions A, C, AeqC, and D. The simulation results are shown as contour lines.}
For the data analysis of \ac{CHAOS}'s stratospheric balloon flight, strong cuts were defined for the individual detector signals in order to contain clean data sets and to filter out noise effects. For the analysis of the measured particles, only those events were analyzed in which the particles triggered the two \acfp{SSD} A and C and the scintillation detector D under the conditions defined for the individual detectors. Figures \ref{fig:data_analysis:fish} and \ref{fig:data_analysis:chaos} show reconstructions of figures \ref{FishSim} and \ref{CDSim} with measured data from \ac{CHAOS}. To compare with the simulation, the simulated data sets have been treated in the same way as the flight data and the contour lines for different count rates have been added to the plot of the flight data. They show where each particle type is expected. Subplots (a) show all data sets that fulfill the conditions just described. For further analysis of the data sets, they were then divided into slow particles that do not trigger the Cherenkov detector (b) and fast particles that do trigger the Cherenkov detector (c) and analyzed separately. Below the two-dimensional histograms, a one-dimensional histogram of the energy deposition in the scintillation detector D is shown.}
\newpage
\change[5]{It can be observed that the tracks of heavy particles, such as protons and helium, can be well separated if only the events where the Cherenkov detector was not triggered are considered. The heavy nuclei are not fast enough to exceed the threshold speed of the Cherenkov detector. Looking at the events where the Cherenkov detector was triggered, we can see that they match the electrons and muons from the simulations. These lighter particles exceed the threshold speed and trigger the detector. It can therefore be said that objectives 1.1 and 2.1 have been achieved. The Cherenkov detector can be used to separate light and heavy particles. Furthermore, the different types of particles within the \acp{GCR} have been identified. However, features remain in the helium regions of figures \ref{fig:data_analysis:fish_Bhigher20} and \ref{fig:data_analysis:chaos_Bhigher20}. These are helium nuclei with enough energy to trigger the Cherenkov detector. They show that the separation between the electrons and the heavier ions is only possible for certain energy ranges. This limitation of the Cherenkov detector must be taken into account when using it.}
\change[5]{Objective 1.2 is met as well. The successful implementation and execution of the \ac{CHAOS} experiment gave useful insight for the further development of the mechanical design and electronics of the \ac{AHEPaM}. For example, new ideas arose to minimize electronical noise in the detectors. Furthermore, the insights from \ac{CHAOS} can be used to improve future \ac{BEXUS} missions from Kiel University. Everything considered, the \ac{CHAOS} experiment can be considered a success.}
% \item Hast du min(A,C) geplotted oder doch max()?\\
% Die deponierten Energien sind statistisch Landau-verteilt. Um nicht statistisch eine Energie zu nehmen, die im langen Landau-Schwanz liegt, nehme ich das Minimum von A und C, wenn beide Werte über den Triggersschwellen liegen.
% \item M\"usste nicht ein Plot Fish-Plot ein negiertes B im Titel haben?//
% Ja, da hst du Recht. Das habe ich wohl vergessen, anzupassen. Ich lasse die Titel weg und schreibe die Triggerbedingung in die Figurebeschreibung.
\change[5]{The results from \ac{CHAOS}'s balloon flight will be used for a total of three Master's theses. The first of the three theses will perform further investigations of the Regener-Pfotzer Maximum regarding the energy-dependence and particle type. In addition to the solar modulation, the latitudinal modulation of the RP-Maximum will be investigated as well. Furthermore, more work on the simulations will be done as part of another ongoing Master's thesis. It will determined how many particles of which type and energy lie in each region of the plots in figures \ref{FishSim} and \ref{CDSim}. For this task the simulated Cherenkov detector will be used. The simulation results can then be compared to the real measurements. In addition, they will make it possible to convert the counts of the real measurements to a flux, which can be compared to measurements of other particle telescopes. To evaluate the flight data, the data must be be cleaned and the instrument properly calibrated. This is the aim of a third Master's thesis.\\
Further ground measurements with the \ac{CHAOS} instrument shall be performed at Kiel University to better understand its behavior. \ac{CHAOS} will also not be the last contribution of Kiel University to the \ac{BEXUS} programme. A new team of students proposed the \acf{SETH} and got accepted to the next cycle of the \ac{BEXUS} programme. Their balloon flight is scheduled for the fall of 2025. \ac{SETH} will also lay important groundwork for the \ac{AHEPaM} just as \ac{CHAOS}. Moreover, \ac{SETH} is the god of chaos in Egyptian mythology.}
A copy of the Campaign Report \replaced[5]{should}{will be} included in this section. \change[5]{Since no Campaign Report was provided on the Alfresco team site up until the deadline for the final \ac{SED} version, the report is not included.}
\caption{Team \ac{CHAOS} in front of the BX35 gondola with the final \ac{CHAOS} experiment. From left to right, back row: Pierre, Myrdin, Clara, Ava, Hannes, Nicolas. Front row: Justus, Lars, Sophie. Not in the picture: Jasper, Janna.}
\item To work effectively and in a goal-oriented way, it is essential to have designated personnel to coordinate the communication between the various parts of the team.
\item Assigning specific roles to certain team members is not an easy tasks. On one hand, the team members doing their bachelor's degree have no experience yet in our field and do not know where their strengths lie. The more experienced team members on the other hand cover tasks from more areas than they are assigned to.
\change{\item The project planning is an ongoing task throughout the whole project. Unexpected issues make it necessary to constantly adapt the original plan.}
\change{\item Deciding on a level of detail for the project planning is not an easy task. Micromanagement can harm the project.}
\change{\item It is very hard to properly estimate the workload of a task in advance. If one has no prior experiences, the workload of a task is easily underestimated.}
\change{\item Deadlines within the team should always be planned with some additional buffer time.}
\change[3]{\item Design choices and plans can change very frequently which makes it hard to keep track.}
\change[3]{\item It is important to stay calm and relaxed even in stressful project phases. This helps the personal well-being and the overall project.}
\change[4]{\item It is important to have enough team members available in critical phases of the project. This is why the identification of these critical phases is so important. For example, a lot of team members were on vacation prior to the \ac{EAR} limiting the amount of work which could be done.}
