# NTFI報告
## 水下聲學、人類活動對於水下生物與環境的影響
* 水下聲學的介紹（在過去海洋、地球科學的應用）
* 實驗報告：水下聲學生物的紀錄、人類的變化、人類與生物的影響（天氣、地震對於生物環境的影響，賭運氣）
* 延伸探討：船隻、海上工程、離岸風場對於海洋生物的影響，融入台灣現行法規與情形）
* 未來展望：海洋聲學對於生物保護的應用
* 附註資料：天氣環境、地點、生態環境、潮汐、時間、儀器參數
* key points
* let them know that Ao is important
* let them thoroughly understand AO as they may not be familiar with this topic
* eff. digit
* 到時候檔案會很大
* 實驗基本部分：儀器、地點、實驗時間、一堆照片
==GPS 25.1445299, 121.8050581==

**==研究筆記[點我](https://hackmd.io/@Y6vFtT4SScOQdqSKfgJfEg/HJ6TUs53q/edit)==**
**==純講稿[點我](https://hackmd.io/AFERx4HjQD-yBDjrhH8fzQ)==**
**==[PPT 共編連結](https://1drv.ms/p/s!AiPxlN1a7YpOhm9PizGLKpktS-0s?e=qE8r70)==**
# **NEW!!!**[ppt編輯連結](https://1drv.ms/p/s!AiPxlN1a7YpOhnXyUPmEvwkE9mB0)

# [NTFI 最新講稿大彙整連結0808](/PPdcIKseQKm3SsFV0HirBA)

:::danger
## Outline
### 前言
* 水下聲學的介紹(在過去的應用、原理...)
* 實驗設計與紀錄(水聽器佈放過程、)
### 研究目的
1. The behavior and activeness of creatures in the region with respect to time and other marine factors.

2. Identifying the characteristics of anthropogenic sounds in water and its changes with respect to time.

3. How human activities affect the behavioral models of marine lifes.`
### 資料分析
* 我們跑出來的圖表意義(spectrum, spectro graph, PSD)
* PSD發現的特殊時段(我們從PSD找到的異常)
* 日落日出的兩根黃色（溫度跟光度跟潮汐）
* 日夜大規模的比較（各頻率、各能量）
* 18:00~19:30 anomaly（鰕虎、雀鯛跟低頻槍蝦的互動）：2.7kHz低、300Hz高（播放槍蝦snapping Shrimp跟鰕虎goby 的聲音、雀鯛的聲音），以及其他在八點左右之後的相似patterns.
* 04:00(船嚇跑了槍蝦和魚，播船的聲音、分析p95在低頻的分貝強度是否比一般情況小)
* 船的活動18:51 00:18、潛水客或其他原因(小尺度，播潛水客的聲音)

* 熱帶海域生態系統的活動狀況及人類對頻譜圖的影響:arrow_right:如何減輕人類對於生態系統的影響
* 生物活躍時間的管制，船舶進港、夜潛、夜晚光害
* 結論
#### 分辨生物的聲音（頻譜、頻率、實際聲音）

#### 聽出人類的聲音活動
#### 生物活動上的習性
#### 人類對於生物的影響
#### 物理因素對於生物活動的影響（日出日落時間還是溫度）

### 提出解決方案

* 在生物活躍時的管制方案
* 生物活躍的因素是甚麼才能應用於全世界

---
:::

## 紀錄
13:12 many people
19:09 3 fish

人聲 0~700Hz
魚 0~300Hz
船 寬頻
>中位數圖

pamguide
series and number
頻譜圖
統計圖
p50 會存到mat檔
makemat.m
cd setting
p50圖 比較好看時間序列上的事件

**[ppt編輯連結](https://1drv.ms/p/s!AiPxlN1a7YpOhm01-xZtvmJF8cBj?e=p2eosN)**

## 報告開始
>前言廢話:<
Greetings everyone,
this is the national team from Taiwan.
Today we're are going to take you to Keelung,
a coastal city that boasts several international harbors,
to conduct our field investigation on acoustical oceanography.

:::success
**topic slide: PART 0 PREAMBLE**
:::

自古至今，人類的好奇心不斷驅使著，大地，太空與宇宙，人類向外探索更多新的區域，然而，驀然回首 on spur of moment,turn back，仍有廣大的未知領域，深埋在大地之下，那正是地球中的內太空，海洋。

台灣四周環海，然而大部分的台灣人對於海卻十分陌生，除了過去歷史上的法律禁令外，若要更深入的了解海洋，多需要較為昂貴的設備。因此，在此次的野外考察中，我們將透過較為簡易的裝置，利用偵測 underwater ambient sound來了解難以觀測的水下環境。

聲學是個在地球科學中新興的研究方式，以電磁波探測地球已是十分熟悉的方式，不過在海洋下，電磁波在海水中的能量隨距離衰減快，因此無法透過電磁波進行探測，但若是透過聲波，則能夠更清晰的看見海洋的地質與生物交流，更能透過水下聲學，反演天氣環境的變遷與聲源特性，除此以外，也能應用於地球科學的其他領域，For example, 可以透過聲學探測行星的內部地質構造，透或是如Acoustic Thermometry of Ocean Climate的氣候變遷觀測研究，甚至能應用於星震學Asteroseismology中，將恆星震動的spectrum 轉換為聲音，能更容易聽見恆星活動的特徵。

>From the beginning of human civilization,
our curiosity has driven us to go far and beyond.
The lithosphere.
The ocean.
The universe.
We had been trying our best efforts to explore the unknown.

>When we slow down our paces
on the spur of the moment
and think about the insignificance of human beings,
we would find even more unknown domains lying silently under the sea’s surface.
That is the ocean,
the inner space that acted as the cradle of lives.
{{35s}}

>Taiwan is an island located on the western Pacific. However, most Taiwanese have little knowledge about the ocean.
It is a common bias that oceanographic investigations take up much effort and funds.
Rather than conducting difficult experiments,
we choose to detect the ambient sound underwater via simple devices
in order to understand the hard-to-investigate marine environment.
Acoustics is a novel research method of Earth science. Although electromagnetic waves are a widely-used way in all spheres of Earth,
the energy of electromagnetic waves declines rapidly with distance in the water.

>As a consequence, the inner space can barely be detected through electromagnetic waves.
Nevertheless, we can see the geological and biological interaction in the ocean
in a clearer way by inverting the changes in the weather environment
and even the characteristics of sound sources through acoustic oceanography.
Furthermore, acoustics is far more than just passively receiving sound waves.
For example, we can use acoustics to explore the internal structure of planets,
observe climate change such as the research *Acoustic Thermometry of Ocean Climate,*
and even do asteroseismology – by converting the spectrum of oscillations of stars into sound,
with the aim of making the observation of stellar activity easier.
{{89s;134s}}

>切入正題

此次的ＮＴＦＩ我們於基隆進行，其located on 台灣東北角， 在實驗前，我們透過浮潛活動，先透過肉眼進行觀察。對於我們來說，這也是十分新鮮的體驗。下水後，繽紛活躍的hydrosphere映入眼簾，我們在水中觀察水下生態系的活躍與氣候變遷對於環境的影響，聽見眾多魚類發出的溝通聲，同時感受著波浪拍打，與台灣東北角由地形與黑潮形成的湧升流帶來的冰冷海水，這些場景皆是在陸地上，我們所無法體會的。在上岸討論後，我們討論浮潛的觀察，決定此次透過水下聲學研究，人類活動對於水下生物與環境的影響。

我們此次的研究方式使用hydrophone進行被動式的 amibent sound underwater監測，first, 我們嘗試熟悉hydrophone的使用與偵測原理，由於聲波是介質的壓力差產生的能量傳遞， when hydrophone 偵測到水壓改變，所偵測到的信號在轉為電壓變化紀錄，後續再藉由儀器校正為正確的聲音強度，就會呈現在水中記錄到的聲音。接著我們再藉由專業人員的協助施放儀器，後續著手資料的分析，了解人類活動與生物所發出的聲音，並觀察兩者間的相互影響。

在儀器設定上，我們的採樣頻率是96kHz，同時開啟了 High-gain讓聲音的偵測更加敏感，我們將儀器設置於圖上的GPS位置，此處屬於珊瑚礁生態的沙灣，為熱帶魚的生態系統，並且是台灣北部夏日戲水的觀光勝地，假日時聚集了許多的遊客與潛水玩家。

在實驗數據收集足夠後，我們回收hydrophone並泡水、清洗(slowly)，下載資料，透過Matlab進行數據的分析，我們首先分析the midian of the 24 hour sptectrum data. as you can see in the picture,during 19:00 and 05:00 時 ，可以看見明顯較寬頻且能量較高的聲音，~~

:::success
**topic slide: "PART 1: Methods of Research**
:::

==加個實驗步驟==
1. 熟悉儀器使用、原理
2. 程式分析下載
3. 討論結果

>We conducted our NTFI experiment in Keelung,
>which is located on the north-eastern coast of Taiwan.
**>We aim to understand
the anthropogenic and biological sound patterns concerning marine conditions,
and observed the interaction between the two at the same time,
>by passively receiving the ambient sounds underwater
>through a hydrophone on 10:30 on July 22nd to 10:30 on July 23rd.
>At last, we further wishes that our research can be applied to conserve the ocean.**
>First, we got acquainted with the use and detection principles of hydrophones **and HOBO**
>and then installed the instrument**s** with the assistance of experts.
>The hydrophone is able to record pressure changes in water,
> which, by Fourier Transform algorithms,
> can be backstepped to the pressure gradient energy transmitted by the original acoustic waves.
> This allows us to hear and even "see" the sounds in the ocean by spectrums, spectrographs, and other statistical data.

>

>This time, we set the sampling frequency at 96kHz,
> and the pre-amplifier is switched on to make the detection more sensitive.
**> The HOBO can collect data of luminous flux and temperature, and is tied on the hydrophone.**
> The location we place the instrument is shown on the map,
> approximately 50 meters off the shore,
> and the depth is about four meters.
> Surrounded by a sandy bay and plenty of coral reefs underwater,
> it is a good place for tropical fishes and crustaceans to inhabit at.
*字幕進度卡* 4:30

:::success
**topic slide: "PART 2: Results and Discussions**
:::

:::warning
**sub slide: a. The Behavior and Activeness of Creatures in the Region With Respect To Time and Other Marine Factors.**
:::

>After collecting sufficient experimental data,
>we retrieved the hydrophone and do some necessary cleanings.
> Then, we could download the data,
> and analyze them with Matlab, a matrix-processing software,
> and Raven lite, a sound-playing software,
> which can separate sounds of different frequencies.

>Here are the three types of graphs drawn by Matlab
> that can be accessed by the original acoustics data: spectrum, spectrograph, and the power spectrum density (PSD) graph.
> Spectrums show us the total amplitude of each frequency at a certain time.
> Spectrographs combine spectrums of a continuous time
> to demonstrate the changes in amplitude
> of a frequency with respect to time.
> The PSD graph represents a certain percentile in amplitude concerning the frequency of the 24-hour spectrograph data.

>At the beginning, we examined the PSD-50 graph
>that shows the median (the 50th percentile) of 24 hours.
>**[~~PSD-50, 24hr, af~~]**
>On both axes, some anomalies can be found as shown on the screen.

>First, we could observe that
>there are four horizontal bands of yellow.
> These represent a larger amplitude of acoustic waves in frequencies of about 1.6 to 2.2kHz, 2.6 to 4kHz, 5 to 7kHz, and more than 10kHz.
> **[~~PSD-50, 24hr, af~~]**
> Having acknowledged that human activities produce sounds below 1kHz **[RAVEN.WAV (some human activities)1218前10秒]**, damselfishes (Family Pomacentridae) produce only waves below 300Hz **[RAVEN.WAV (some fish)0239(01:30)]**, and boats generate acoustic waves for broad and continuous frequencies simultaneously, **[RAVEN.WAV (some boat activities)0357 02:40後]** these patterns are thought to be the sounds from snapping shrimps (Superfamily Alpheoidea) **[RAVEN.WAV (some snapping shrimps)0603一開始]**. ==As the snapping shrimps generate acoustic waves by releasing bubbles in their joints when predating, a single "snap" should have continuous frequencies as well.== Therefore, we propose that with four snapping patterns present, at least four types of behaviors are occurring in the region. This may mean at least four species of snapping shrimps, or that shrimps of different life stages behave differently. To simplify the discussions, we would refer to the 1.6 to 2.2kHz anomaly as "Pattern A", and sequentially, "Pattern D" for the stronger part of more than 10 kHz. We would discuss the changes in behavior patterns with respect to time more thoroughly in the following sections.

>Next, two significant yellowish vertical bands are observed at 18:47 through 19:18 on July 22nd and 04:41 through 05:17 on July 23rd. This indicates a 10dB-some increase in amplitude at all frequencies compared to the background. **[~~PSD-50, 1800-2000, af~~],[PSD-50, 0400-0600, af]** After checking the changes in luminous flux and temperature collected by HOBO, as well as the weather statics by the Central Weather Bureau, **[PSD-50, 1800-2000, af]**

---
],[HOBO-luminousflux, 1800-2000],[tides(CWB), 1800+1900+2000],[wind(CWB), 0800-2000]**
>**[PSD-50, 0400-0600, af],[HOBO-luminousflux, 0400-0600],[tides(CWB), 0400+0500+0600],
>[wind(CWB), 0400-0600]**
>we found that the time of increase in luminous flux fits most with the time of increase in amplitude. **[PSD-50, 1800-2000, af],[HOBO-luminousflux, 1800-2000],[PSD-50, 0400-0600, af],[HOBO-luminousflux, 0400-0600]** ==Further more, we found that some snapping shrimps can sensor the alternation in luminous flux.== To conclude, we consider these two phenomena to be the results of increased activities from marine species like damselfishes and snapping shrimps of all Patterns A to D.

>To move on, we examine the differences in general patterns between day and night. During the day, a relatively strong amplitude occurs at 200 to 500Hz randomly; while at night, some energy is comparably higher at 1.8 to 30kHz. **[spectrum, 1200, af][spectrum, 1900, af][spectrum, 0000, af][spectrum, 0500, af]** After listening to the sounds by using Raven Lite, we found that the 200 to 500 Hz spikes occurring in the day was generated by bubbles and paddling from snorkellers, scuba divers and free-divers. **[RAVEN.WAV, a bubble, 0.2~0.5 kHz]** The difference in higher frequency from 600 to 30000Hz, on the other hand, might be affected by the nocturnal preferences of snapping shrimps of Patterns A to D. However, we found a relatively strong activity density from snapping shrimps C (5~7kHz), which we think is related to the changes of tide levels. **[PSD-50, 1030/0722, af][PSD-50, 1030/0723, af]vs[潮位表(CWB), 1030/0722][潮位表(CWB), 1030/0723]** We also found that this type of phenomena occurred in other researches, such as the ones conducted by professor Huang. **[PSD-50, 黃千芬, af]vs[潮位表(CWB)]** All previously mentioned PSD graphs represents a larger amplitude in the daytime when the tide levels are higher, so we infer that the snapping shrimps of Pattern C (5~7kHz) has higher activeness during the day when relative high tides are present.

>Thereafter, we examine the anomalies on smaller scales. A spot of blue from 18:32 to 19:12 demonstrated a decline in 1.6 to 4 kHz but an increase in less than 300Hz. **[PSD-50, 1800-1930, 0-5kHz]** We also checked the PSD-5 and PSD-95 graphs, which turns out to have the same altering patterns. **[PSD-50, 1800-1930, 0-5kHz][PSD-8, 1800-1930, 0-5kHz][PSD-95, 1800-1930, 0-5kHz]** It means that the energy is declined or gained generally, but not by a ream of human activities that influenced the median. The 1.6 to 4kHz anomaly thus indicates fewer activities of snapping shrimp in Pattern A and B, and the 300Hz rise represents higher activeness of damselfishes. After searching for papers on the Internet, we proposed that the patterns may be associated with the interactions among snapping shrimps, damselfishes, and gobies (Family Gobiidae)- a species that are barely found by passive acoustic methods but is investigated existing in the regions of our field investigations. **[RAVEN.WAV 1957(01:20~01:40)]** July is thought to be the peak of the reproductive period for damfishes in which they generate louder acoustic waves at night and show strong territoriality. As Gobies and snapping shrimps are in mutualism on avoiding danger, we think that the loud sounds of damselfishes lead to the alert of gobies, which further prevents the snapping shrimps from leaving their caves. Similar patterns that snapping shrimps tend to hide when a higher amplitude of damselfish noise is present can be found several times at night, this also acts as a confirmation of our theory.

:::warning
**sub slide: b. Identifying The Characteristics of Anthropogenic Sounds in Water and its Changes With Respect To Time**
:::

>When we zoom in to even smaller scales of minutes, Human activities act as another important sound source, too. Some random vertical lines of smaller amplitudes in all frequencies are generated by boat propellers **[RAVEN.WAV, some boats]** Also, during the day, we could find the bubbles of breathing from scuba divers and free-divers at a rate of around 10 to 20 times per minutes. Additionally, breathing sounds made by the snorkellers are also easy to find. **[RAVEN.WAV, 0849, <500Hz]**

:::warning
**sub slide: c. How Human Activities Affect the Behavioral Models of Marine Lifes**
:::

>At 04:00, a sudden drop in amplitude of all frequencies occurred. The amplitudes gradually recovered to normal in the following 15 minutes. **[PSD-50, 0330-0430, af]** By Raven Lite analysis, we heard an extremely loud noise of a boat. **[RAVEN.WAV, 0330-0430]** We suspected that the pattern was made by the hiding of species as boats passes.

>As the spot was also open to visitors and divers, frequent anthropogenic noises can be found. They (<500Hz) share the same frequency bands with the damselfishes (<300Hz), but the sounds are relatively louder. This may interfere with the communications and courtshipping of damselfishes. Additionally, damselfishes are diurnal-preference species; however, we didn't see significant pattern differences as the snapping shrimps' does. This may act as evidence of how human activities alter the behaviors of other creatures, which, eventually, disrupt the whole ecosystem.

## PSD發現的特殊時段
從這張圖中我們發現到了以下幾點：
首先，我們可以看到在圖中有兩個偏黃色的部分，代表當時各頻率聲波振幅較大。在確認過hobo的資料後，我們發現該時段有激烈的光線強度變化，代表當時正值日出與日落，我們認為可能是日夜交替時生物感應到光度的變化，使得其活動力變強，產生了強度較高的聲音。

接著，我們比較生物在日間與夜間的時頻譜能量差異。我們發現低頻(0.2 ~ 0.5kHz)在日間有較多能量較高處，而夜間也有幾處能量較高，我們經過分析之後認為日間的低頻能量主要來自人類潛水時發出的聲音，夜間則是魚類的叫聲。而在較高頻(0.6kHz ~ 30kHz左右)，則是夜間有較高的能量，我們認為是夜間槍蝦活動較強所造成的聲響。此外，我們發現頻率為5~7kHz的聲音在早上強度較大，晚上強度較小，經過與當日的潮位比較之後，我們認為潮位較高時，槍蝦也較多，同時我們利用黃千芬教授的分析資料，發現在潮位越高的白天時段，槍蝦的活動聲音也就越大，由此我們可以推測潮位會影響槍蝦的活動，且潮位越高，槍蝦的活動力越強。...從這張時頻譜圖與當時的潮位比較可以發現，在潮位較高的時間槍蝦的活動力較強，由此我們可以推測潮位會影響槍蝦的活動，且潮位越高，槍蝦的活動力越高。

再來，我們可以看到在18:00~19:30時有anomolies，在2.7kHz有較低的能量，而在0.3kHz則有較高的能量，我們推測這與鰕虎、雀鯛和槍蝦的交互作用有關。因為7月正是雀鯛的繁殖季節，==雄性雀鯛會發出強烈的聲波求偶==，雀鯛的強烈聲波引起蝦虎躲避，和蝦虎魚有共生關係的槍蝦也跟著躲避，導致其聲音強度減弱。
從夜間低頻聲音強度較高的時間也可以看到槍蝦的聲音明顯變小，這個結果證實了我們的猜想。

另外，我們發現在04：00時有一段時間個頻率的聲音強度皆有減弱，在利用Raven lite進行分析之後，我們聽到當時有船的引擎聲，我們推測船經過所發出的聲響會使魚群和槍蝦躲避，或是船的聲音干擾到生物，導致發出的聲響變小。等到一段時間船走了之後，生物的聲音又漸漸回復成較大聲的狀態。

最後，在小尺度的時間方面，我們也發現在部分時間有潛水客水肺潛水時所吐泡泡的聲音，頻率大致和魚類相同，可能會成為魚類溝通的噪音，進而導致魚類的行為受到影響。例如：雀鯛原本為日行性動物，但是在人類的干擾之後變成日夜間有活動，如此現象可能會對魚類的身體狀況產生影響。

## 結論

:::success
**topic slide: "PART 3: Conclusions**
:::

>From the PSD graphs, we can find, listed in the order of commonness, snapping shrimps, damfishes, boats, and sounds of diving activities.

>==這篇是講稿，PPT放表就好==For marine species, we found 4 behavior patterns of snapping shrimps and at least one species of damselfish present in the area. Their activeness is related tightly to luminous flux. These are the findings that we observe in the graphs. As for snapping shrimps, diving and swimming sounds seem not to influence their activities. They are most energetic at crepuscular times, and Patterns B, C, and D have a slight nocturnal preference as well. Pattern A and B have a drop in amplitude at sunset, which we thought to be evidence of its mutualism with gobies. Pattern C is more active at higher tide levels during the day, while the others show no preferences.

| Patterns | Snapping shrimps A | Snapping shrimps B | Snapping shrimps C | Snapping shrimps D | human activities |
| -------------------------- | --------------------------------------- | --------------------------------- | --------------------------------------- | --------------------------------------- | ---------------------------- |
| | | | | | |
| Frequency | 1.6~2.2kHz | 2.6~4.0kHz | 5.0~7.0kHz | >10.0kHz | <500Hz |
| | | | | | |
| Sunrise (04:30~05:30) | 77dB | 80dB | 83dB | 85dB at most | 75dB |
| Amplitude at sunrise | Very High | Very High | Very High | Very High | Medium |
| Possible Reason | crepuscular preference | crepuscular preference | crepuscular preference | crepuscular preference | |
| | | | | | |
| Daytime (06:00~17:30) | 72dB | 72dB | 74dB | 71dB at most | 89dB |
| Amplitude during daytime | Low | Low | Low | Low | High |
| | | | | | |
| Sunset (18:45~19:15) | 65dB | 60dB | 85dB | 86dB at most | 81dB |
| Amplitude at sunrise | Very Low | Very Low | Very High | Very High | Medium |
| Possible Reason | more active damselfishes | more active damselfishes | crepuscular preference | crepuscular preference | |
| Infers that | in mutualism system with gobies | in mutualism system with gobies | **NOT** in mutualism system with gobies | **NOT** in mutualism system with gobies | |
| | | | | | |
| Night (20:00~04:00) | 73dB | 76dB | 77dB | 75dB at most | 61dB |
| Amplitude at sunrise | Low | Higher | High | High | Low |
| Possible Reason | **NO** significant nocturnal preference | nocturnal preference than diurnal | nocturnal preference than diurnal | nocturnal preference than diurnal | fewer divers and snorkellers |
| | | | | | |
| 04:00 massive boat passing | 67dB
decrease by |6dB|4dB|10dB|10dB
Others|||more active in the daytime of higher tide levels

>As for damselfishes, they generate acoustic waves of less than 300Hz. While we had found data that they are diurnal species, we see no difference in their activities between day and night. This is considered strange, and we thought that daytime human activities such as diving and snorkeling are the reasons for this phenomenon. Additionally, divers also generate waves of approximately the same frequency band, which may influence the communication and courtshipping of damselfishes.

>Ship propellers, on the other hand, produce broad-frequency noises. In PSD graphs, although they are weaker sound sources, they are commonly existing as the median amplitude shows a drop. These sounds may make other marine species harder to communicate with one another as the background is noisier.
>Moreover, with the two main creatures both being sensitive to changes in light, artificial light sources may also affect the marine ecosystem here.

在經過 spectrum, spectrograph, and the power spectrum density (PSD) graph和海洋聲音的分析之後，我們做出了以下結論：
1.根據PSD圖和聲音分析，我們認為海洋生物方面至少有四種槍蝦以及魚類的聲音，槍蝦的聲音分布在1.6kHz以上，魚類的聲音則在約0.3kHz左右。我們發現在光強度變化大時，也就是日出以及日落時，海洋生物的活動力較強，在光強度較弱時，也就是夜間，海洋生物的活動力次強，在光強度較強時，也就是白天時則最弱，另外，潮位越高，C行槍蝦的活動力也較強。
2.人類對海洋形成的噪音主要有潛水時的泡泡聲和船舶的引擎聲，潛水泡泡聲主要在0.5kHz以下，且主要分布在日間，船舶的引擎聲則較寬頻，且在全天皆有零星分布。
3.關於人類行為對海洋生物的影響，我們認為日間人類潛水產生的泡泡聲，因頻率大約與魚類的聲音重疊，可能會影響魚類的溝通，而我們推測整天皆不時出現的寬頻船舶引擎聲，可能會將魚類以及槍蝦嚇跑，或是阻礙他們的溝通。另外，人類行為所產生的光強度改變也有可能會影響生物的行動。

:::success
**topic slide: "PART 4: Solutions?"**
:::

>With the possible influences of human activities to marine species shown, what can we do to solve the issues?

>We have acknowledged that light and anthropogenic sounds affects the behaviors of marine species. If we cannot ban marine traffic and diving, at least some regulations, *inter alia*, not to dive during sunrise and sunset, or reduce the time and strength of ship illumination, can be made to reduce the influences of human activities on marine species.

>However, although models are proposed in our investigation, the model may have a slight error when implementing on coral reef ecosystems all over the world, as maybe the system may be controlled by more factors unknown. With the aim of applying this result worldwide, further investigations on differing regional ecosystem components and behaviors may be done. By then, suitable conservation methods can be designed respectively, and do contributions to the sustainability of the oceans.

## 提出解決方案

海洋中的魚蝦除了會因為光度、潮汐的變化導至活動力變動之外，也有可能被其他生物以及人類的聲音影響，若人類活動產生噪音或是亮度變化，甚至是其他因素造成某種生物的行為受到影響，其他生物的活動也可能產生連鎖效應，最後影響到整個生態系。因此找出人類可能干擾海洋生態的方式，並將這些干擾降到最低，是我們現在必須做的。

目前我們已知光的亮度變化和人類噪音會影響到海洋生物的行為，如果我們沒有辦法停止船舶航行、潛水、會產生光亮度變化的各種活動，那麼我們可以嘗試管制人類的活動，利用政府與民間共同努力，憑藉立法訂定規範以及日常對民眾宣達關於海洋生學的知識來達到目的。尤其是在生物活動強烈的時段，訂定規定以減少對海洋生物的干擾，相信對整個海洋生態或許會有正面的影響。

但是，我們研究僅限於基隆海域，並不一定每個海域的生物活動力都是受到相同因素影響，所以應該要擴大研究範圍，了解每個區域生物的習性與影響其活動能力的因素，並依照不同地區海洋生物的習性制訂保護海洋生物的規範，以保護海洋環境的永續性。

~~18:39~19:09時在約2.2~3.8kHz處有一個振幅較小值，我們猜測可能是某一種槍蝦的主要聲音頻率為3kiloHz左右且聲音較小，在該段時間活動力高，製造了較多分貝低的聲音，低分貝聲音占比變高導致中位數降低。為了驗證我們的猜想，我們利用Matlab跑出了在P?andP?的時頻譜圖，從這張圖我們可以得知……

~~Second, 我們發現在19:54 to 20:15 and 20:24 to 20:33 and 20：52 and 21：06 and 21：15 and 21：27 and 22：33 all frequency 有船的引擎在打水的聲音
1957 1:20秒有一連串類似放屁的聲因(? 聽起來像比較扁的泡泡破掉聲。同時我們也發現，在2kHz的中位數振幅有明顯的下降。

~~third, 在00：15~03：24 all frequency 有 (boats？？？)
==0018.40~0019.20 boat confirmed==~~

~~Forth, at 03：57 , all frequency 有超級明顯的黃色轉藍色，代表當時個頻率的聲音明顯變弱，我們推測可能是輪船經過，導致魚群被船隻給嚇跑或是被輪船的訊號所干擾。 ~~

~~Fifth, at 08：15, PSD呈現40kHz脈衝訊號~~

>As you can see in the picture, during ? and ? , a distinct high-energy spectrum(高能量頻譜) can be seen.


## 0728修改
1 標題：題目太廣，分析的東西太特定?
==論(...)和人類活動對於槍蝦與雀鯛的關係==
生物的活動特性in基隆望海巷
，一開始先把題目講清楚
>廢話真的可以不用，包括影片
>
->abstract(1min),
>直接放圖，引起注意

motivation(2min),
>廢話省略，實驗原理減少

experiment(7min)
>省略過多推論過程，甚至直接講結果
>
->
取出各種生物活動頻段
天（日夜生物活動，先日再夜，光度&生物相關係數 or 光度衰減率&生物相關係數，溫度&生物相關係數）
->小時（雀槍虎的相互作用，三者聲音相關係數）
->分鐘（人類活動，人類活動&生物活動相關係數）
conclusion(2min)
->三個東西
the other (2min)
圖形 colorbar沒有呈現好(特徵不清楚，對比?)->colorbar(jet)
圖和圖示都不夠清楚:(
量化人類活動？
>將槍蝦的聲音當成背景雜訊濾除，凸顯低頻的魚類活動與人類活
動。

新講稿（縮短時間）

海洋生物的照片?

實驗進行過程照片?

海洋資源? connecting to other factors?

寫下屬於自己的動機

未來展望：地面天氣的準確性？

PSD是什麼？講清楚
常用QA投影片
槍蝦發聲的物理原理(假設只有一種槍蝦) #孔蝕噪音capitation noise

槍蝦發聲的強弱代表(可以從發聲原理推)

reference 翻譯

Conclusion 太多頁

首尾呼應

summary(一定只能有一頁)

>what you did great in the journy?
>what you can make it better?
>the title of the slide should match the graph shown below.
>what is the most important thing should present in the NTFI

P3 Most Taiwanese have little knowledge about the ocean.？？？

acknowledgement 拉到前面

### 技術分析
相關係數about 槍蝦a, b, c to 潮位高度、溫度、光度...去證明2-a的各項數據從統計資料上是有相關的，再對到圖表上證明。

Anova 分析 : 日夜行性槍蝦對特定頻率的分貝平均值。(to support the evidence of snapping shrimps activeness.)

## 0730修改

### 別人的問題
* 圖片alignment
* title 和內文沒有對到
* 槍蝦為什麼要發聲(為了嚇阻、獵食)，逼不得已才會嚇阻，一般來講是能躲就躲，所以躲起來的時候不發出聲音是合理的。
* 科學方法直接炸掉
* do we need wenze curve？
* 字太多、太密
* 單位：dB、聲強度、統計數據
* 槍蝦是叫更大聲還是更多隻在叫？

# 我們的問題
* ## 針對投影片呈現
- [x] 頭往右移，左邊標上國籍和IESO NTFI字樣(國旗放中間?)
- [x] 給個短小吸引人的主標題，原本的標題作副標
解決方式：改為短小的主標題
- [x] ABCD 加上頻率域(ABCD倒轉，FREQUENCY BAND IN ALL GRAPHS)
- [x] 圖表橫軸調整為四個時序的時段比較清楚(以透明色塊加底)
- [x] 縱軸的範圍可能要調整成讓ABCD的趨勢更明顯
- [x] p.15 habit-> biological clock pattern, activity time
- [x] 系統圖->釐清所有因素的交互關係
- [x] acknowledgement 放後面(？
- [x] SSOM死去，請精簡化
轉為未來展望，以文字敘述即可
- [x] "In chinese"

* ## 針對研究背景
- [ ] 把槍蝦聲音=活動行為 的假說做好
- [ ] https://www.science.org/content/article/snapping-shrimp-close-their-claws-so-quickly-they-create-shock-waves-video-reveals-how
- [x] 時間地點
- [x] 布放儀器的地點圖(全景照片)讓觀眾身歷其境
- [ ] 辨認生物的過程（文獻背景）、三種生物的聲紋與頻譜跟頻段
- [x] background and settings 要多補充，定義要明確．
- [ ] 引注要在內文裡標注起來
- [ ] 槍蝦住在洞裡，對人類活動的反應比較沒有影響(？．
- [ ] 會叫的雀鯛蠻少的，三點白的雀鯛跟網文雀鯛比較會叫？）
* 三點白雀鯛：[台灣野生動物之聲資訊網](http://www.taisong.org/search-view.php?ix=1&nc=382363&v=s)
* 網紋雀鯛：[Mark McGrouther(2021),Headband Humbug, Dascyllus reticulatus (Richardson, 1846),Australian Muesum.](https://australian.museum/learn/animals/fishes/reticulated-dascyllus-dascyllus-reticulatus-richardson-1846/)

* ## 針對數據
- [x] ~~槍蝦在船經過幾分鐘後才不見?~~
- [ ] data interval
- [x] 收音水平範圍、推算儀器極限、收到的槍蝦聲源距離不同導致能量衰減
- [x] 其他頻段：蝦虎0-200Hz，雀鯛300-800Hz，人？

* ## 針對統計資料呈現
- [x] "biological habit->Active times"
- [x] 魚活動增加後會影響到槍蝦的pattern?
- [x] data interval (評審問到再講)
- [ ] 加強PSD的解釋、意義
- [x] Pattern 趨勢 -> Band width 頻率區
- [ ] 4 band 分類的學理根據、槍蝦的發聲原理、何種行為模式、報告？
* 群體差異假說、性別假說
- [ ] 統計時丟進去跑的單位跟數據解釋
- [x] ANOVA A-D 為什麼用平均不用中位數？應該要用中位數待修正
- [x] pattern A-D的差異在哪，用什麼方式證明四種區段有所差異(p.8配ABCD的時頻譜)
- [x] 人跟魚的關係比較直接重疊，會被魚干擾導致相關係數不對，建議以人類活動的週期反推頻率段的變化趨勢，再取頻段進行分析
- [ ] 講清楚4 band 的週期性與定義（基本上就是從功率頻譜圖觀察分析得來分辨四個不同區段的上下界）
- [x] 證明abcd不相似．
- [ ] ==把統計的方式敘述好，讓統計可信度出來，就差不多可可了==
Patty's comments :
- [x] 1. correlation coefficient 可能要再確認一下
- [x] 2. coefficient 位數不用這樣多
- [x] 3. 先從DATA 計算出的PSD 定義
- [x] 4. 在PSD 你們歸納了哪些頻段 屬於哪些訊號 （如果有reference 佐證會更好）
- [x] 5. 溫度的變化不大 可能造成相關係數不明顯
https://www.researchgate.net/publication/258673432_Seawater_Temperature_and_Wind_Speed_Dependences_and_Diurnal_Variation_of_Ambient_Noise_at_the_Snapping_Shrimp_Colony_in_Shallow_Water_of_Southern_Sea_of_Korea 溫度與槍蝦關係相關文獻
* ## 針對研究賣點
- [x] 專注於人類活動--海洋噪音議題＆航道噪音
- [ ] SSOM死去，請精簡化

### 會議資料如下:
:::spoiler
## 會議記錄

### 7/30 google meet

林卉婷下午1:12
"Acknowledgement" page
林卉婷下午1:17
https://www.science.org/content/article/snapping-shrimp-close-their-claws-so-quickly-they-create-shock-waves-video-reveals-how
黃千芬下午1:35
Sorry that i need to move to the next meeting. Might log in again during the drive
PeiYing Patty Lin下午1:49
因為溫度變化不大啊
林卉婷下午1:55
"Active times"
PeiYing Patty Lin下午2:08
Patty's comments :
1. correlation coefficient 可能要再確認一下
2. coefficient 位數不用這樣多
3. 先從DATA 計算出的PSD 定義
4. 在PSD 你們歸納了哪些頻段 屬於哪些訊號 （如果有reference 佐證會更好）
林卉婷下午2:14
Tina建議：
1. 利用ANOVA進行的分析跟結論也是要確認，因為你用的4個pattern，df=3，總覺得怪怪的
2. PSD with time 去跟其他時間點去比較 correlation好或不好，這部分請看Patty的建議，因為這個是接到你們結論的地方
我先下線喔！同學們，請繼續加油！
台灣的海洋真的很需要你們喔～
PeiYing Patty Lin下午2:17
In chinese就好

### 8/2 會議記錄
會前：
1. ~~跑各種ANOVA試試~~

會議：
1. ANOVA 與post hoc的組別設定
* ANOVA：檢定ABCD四組槍蝦的差異
* z, median, sigma: 船、晨昏晝夜時段
2. 實驗組別定義釐清
* 雀鯛：300-700Hz-m人
* 人類：800-1200Hz，200-300Hz，跑擬合找幅度係數
* 蝦虎：0-200Hz-m人，100-200Hz-m人
* 潮汐：24個點跑相關
5. PPT呈現建議討論
*

會後：
1. 繼續跑數據
2. PPT實驗結果部分需修改
3. PPT投影片呈現問題修正

:::

## 槍蝦行為模式推導
### [連結點我](https://hackmd.io/AkTs0aY7QJ6NIwoL3mtC4w)

8/11

### 講稿簡化再簡化!!!!!!!!!!!
### 邏輯?
- [ ] motivation 要修，建議把prospect放在前面
- [ ] anthropogenic factors?(hydrophone 只會錄到聲音)
- [ ] backsroung setting (強調旁邊的海科館協助我們進行，潮境公園是生態保育區，gps可省:()
- [ ] p4的cites?(重點:anova的參數界定?psd's characteristics of what we have recorded, )
- [ ] 介紹槍蝦(用圖代表與蝦虎的共生關係)
- [x] p4的band 換成 frequency section
- [ ] p4 直接定義所有會用到的頻段，它的範圍及定義
- [x] p5(標題建議可改 four different snapping shrimp sections/various patterns observed in different sections)
- [x] p5的兩個test(p value, 虛無假設為abcd都一樣，p(一樣的機率)<0.001，拒絕虛無假設(or f value > ))
- [ ] p5的anova test within groups可以刪掉，post hoc test標題改掉，判斷是否拒絕虛無假說的邏輯寫下(如何判斷?結果如何顯示?)
- [x] p5接到p6有點突兀?(解釋abcd的數據意義，如數值高表示聲音強度高，並且同樣在晨昏時有peak,為什麼會這樣呢?)
- [ ] p6標題(sounds是對的嗎?)(相關係數表格標題簡化，每個表格後寫結果)(新增小結+假說)
- [ ] p8目的不明?(直觀跟數據不太符合?聲音強度與光度?應改為，day and night 的 pattern符合，但 dawn and dusk不太一樣，所以要進一步解釋)
- [ ] p8 diurnal定義，槍蝦整天都會叫，如何定義日夜行性?
- [ ] p9 標註c、d頻段
- [ ] p11、12人類與雀鯛毫無關聯(整個報告的邏輯順序須重新思考)
- [ ] p13 宜改為interferrence interaction among 3 animals.
- [ ] p14 標題改:silent anomaly
- [ ] p14 在圖上標註船的發生時間
- [ ] p15 結果表格可以省略?
- [x] P16 conclusion 條列式寫，從最直觀的現象、數據推到較深層的含義:

1. 槍蝦在晨昏活動最大，雀鯛、蝦虎基本上也是
2. 船會影響，但人類潛水活動我們並沒有足夠證據證明)
- [x] p17 prospect, 應該要扣回我們最有證據的結果(管制船，而非人類，因為在保育區內。)

### 8/13
#### 講稿問題
- [x] p.5 先引述槍蝦總頻率，根據觀察分mode的依據，描述分頻段的用途(證明這些都是槍蝦的frequncy band)

**[範例投影片呈現](https://www.marineinsight.com/environment/effects-of-noise-pollution-from-ships-on-marine-life/)**
- [x] 國外多主要用條件子句陳述
- [x] 介系詞
#### 投影片問題
- [x] 沒有介紹雀鯛，順便在結論加雀鯛也有觀察到晨昏性，船那邊在更仔細說明一下，人類活動的pattern沒有color bar.

- [x] Prospect 不要放human(因為要講說需要減少人類在晨昏時刻的活動，就算沒有關係，也可能有其他的汙染源影響到槍蝦和其他生物的生存環境)
- [x] p14 Discussion 有點花，改回去+- ㄏㄏ:(
- [x] 讓投影片有空間放頭
- [x] p.2 Scuba Diving 太專業 改成recreational sports in conservation area. (較為廣泛包含所有的水上水下活動)
- [x] p.2 what influences marine species and the ways to protect the life below water. 是指海洋生物有什麼影響 要改意思
- [x] p.3 把音檔的音量調高(注意有時候會沒有聲音)
- [x] P3.應該是 more active the snapping shrimps are （不是activeness ><)
- [x] 記得把滑鼠改成類似投影筆的圖示
- [x] P.4 可以分三個階段的動畫，讓觀眾更清楚我們在講的重點
- [x] p.5 先引述槍蝦總頻率，根據觀察分mode的依據，描述分頻段的用途(證明這些都是槍蝦的frequncy band)
- [x] p.5 在講到時間分段的時候再把時間分段表格放出來即可
- [x] p.6 F-crit
- [x] p.6 Null hypothesis 去匡去標色，匡拒絕虛無假設：they are difference. should emphasize on that every two modes is different, not the null hypothesis. 去除星星或是標星星的意思
- [x] Use p<0.01 or p<0.001 is enough
- [x] p.6 四張圖的標題沒有對齊
- [x] p.7 kHz要連起來
- [x] p.7 decrease的地方要標記得更清楚(should be written down in psd/sound intensity decrease, by the way, the frequency band we emphasized are 1.4-4kHz, not 1.6-4kHz. )
- [x] p.7 標示出雀鯛所在的頻段(add an introduce of danselfish.)
- [x] p.8 大小寫注意一下
- [x] p.10 et al. should be moved down to the lower side, classic human side 的編號改掉(改成B:HUMAN DIVING SOUND,)
- [x] 圖表中的"the"其實都可以去掉。
- [x] p.12 "sudden drops of intensity", k and Hz should not have a space between them.
- [x] p.12 要把低頻率船的聲音標出來，表示出當時有強烈的低頻聲音，也可以考慮播音檔，另外，橢圓框也可以改成方框。
>- [ ] when the amplitude above the highest limitation, the signal in high frequency woule be "clipped", maybe after the ship passing through, the result of the quietness of snapping shrimp is not correct
- [x] 魚類活動比較容易受到船的聲音的影響，槍蝦不一定。
- [x] ship pattern should be put in the left side, and zoom in the lower frequency part(the ship anomaly might originate from instrument mistakes)
>- [ ] ship anomaly 一樣需要遮罩(強調槍蝦的band被船影響)
>- [ ] stars could be put in the median group of ship anomaly in order to show a clear sight of which group would be more influenced(a,b)
- [x] p.12 同一組的長條圖可以放近一點
- [x] biological clock should be deleted in the meddle of the arrows
- [x] word shape(should be turned into Ariel, or......, not times new roman)
- [x] Think more about the interactions among damselfish, gobies, snapping shrimp.
- [x] **modes should be altered into "TYPEs"**
- [x] the frequency bands of each type of snapping shrimp should be added in the conclusion slide.
- [x] 多想
- [x] To emphasize on the way we detected snapping shrimp CAN be used in the ocean worldwide， in prospect(強調方便、範圍大、也可長期測，一方面也可以減少人類親身進入保護區調查所造成的影響).
- [x] the purposes of motivation slide are attracting people to engage in our investigation. It should be designed in more funny way and clearer way so that all the audiences would tend to listening to your presentation.
- [x] 建議放在動機：潮境公園有限制晚上6點之後不能有人類活動(其實是有規範的，可能是已經被限制了所以圖表中比較沒有記錄到人類活動的聲音，他們也是個貢獻)
- [x] 人類活動只有白天可以有，所以這條件可納入報告(相關係數分析)

## Q&A綜合集
1. 為什麼我們用槍蝦?
> 因為槍蝦廣泛分布在珊瑚礁海域，且聲音很大，在國際上已有人提出使用槍蝦當作探測海中環境的主動聲源，而聲波在海中較易傳遞訊息，是個適合用來偵測海洋的好工具。除此之外，也可以減少人為調查行為破壞環境的機率。
2. 研究的特點是什麼?
> 我們發現這份研究方法是利用較為簡單的儀器去探測海洋環境。一方面其優點是可以降低成本、使用性長、可即時回傳資料，大幅降低探測海洋的難度。
3.