Cleveland Bay Automated Marine Weather And Oceanographic Station

This dataset contains meteorological and sea temperature data from two automatic weather stations moored at different times and locations within Cleveland Bay (Townsville).

The first weather station was deployed on Channel Marker #10 (19° 11.95S, 146° 51.89 E). Data from this weather station covers the period 3rd July 1990 to 21st October 1992. This weather station was then removed and redeployed on Cape Cleveland.

The current weather station is located on channel marker #S5, approximately 1 kilometre from Bremner Point on Magnetic Island (19° 09.35S, 146°52.87E). Data collection from this station commenced on the 7th December, 1999.

These data are collected to support scientific research at AIMS. Data are made available on request to other researchers and to the public.

The current weather station is a Campbell Scientific Australia System.

Data recorded: Sea Temperature (1.9m and 8.5m at MSL), Barometric Pressure, Air Temperature, Solar Radiation (PAR), Wind Direction True (scalar averaged), Wind Speed True (10 min average prior to every half hour and maximum wind gust).

1. Operation and Weather Sensors

The weather stations collect and store data in electronic memory every half-hour. A central base station calls each remote station regularly using HF radio or telephone lines. The data is transmitted over the radio as a frequency shift keyed signal, organised as packets of information. Errors are detected using parity and check sum methods. Invalid packets are identified by the Base Station, which requests they be sent again. This concept allows recovery of a very high percentage of the data despite poor communications. Remote stations store data for 21 days. Features such as automatic operation, remote control, remote time setting, built in diagnostics, have been developed and incorporated.

The sensors are a key part of a weather station. The following are chosen considering the cost, reliability and accuracy.
* R.M.Young manufactures the wind sensor, a model number 05103. It is a propeller type with the advantages of being highly linear, highly interchangeable and having a low threshold. Wind direction is measured as the direction the wind is coming from.
* The solar radiation sensor is an Under Water Quantum Sensor made by Licor. It measures light in terms of its "Photosynthetically Active Radiation" (PAR). The spectral response is defined and weighted. Drift due to aging of the filters has proven to be a problem, but this applies to similar units too.
* Temperature sensors are all Omega Interchangeable Thermistors. These are interchangeable and have high accuracy, but reliability has proven a problem. We are considering alternatives.
* The barometric sensor was a modified Aanderaa type on earlier stations. The Mk2 stations were fitted with a Weathertronics Unit. Now all stations are Mk3 stations fitted with a Vaisala barometer which is more interchangeable and more accurate.

2. System Accuracy

System accuracy is calculated as the sum of errors caused by:
* Calibration
* Interchanging sensors
* Drift with time
* Effects of an ambient temperature range from 0-40 degrees C.

The following are the specifications of the sensors used with Mk3 stations. A new sensor suite will be used with Mk5 stations, partly based on the Vaisala WXT510 weather sensor.

Both the temperature and wind sensors are interchangeable, and not individually calibrated, though some individual sensors have been checked against standards.

* Air Temperature: Interchangeable thermistor and electronics is within +/- 0.4 deg. C, with a 30 seconds settling time in air. There are additional errors due to the aspiration of the temperature screen at low wind speeds.
* Water temperature: Interchangeable thermistor and electronics is within +/- 0.4 deg. C, with a 30 minutes settling time in water. A higher precision in situ calibration is normally used (around +/- 0.1 degrees), traceable to a 0.04 degrees standard.
* Solar radiation (PAR): +/- 5% of reading. Sensor drift is approximately -4% per year initially.
* Barometric pressure: +/- 1 hecto Pascal.
* Wind speed: 2% of reading +/- 0.1% FSD.
* Wind direction: 2% of reading +/- 0.1% FSD.

Electrical settling time for solar radiation and wind parameters is 7 seconds. This is necessary for anti-aliasing filters. Mk1 and Mk2 stations averaged 16 samples over the 16 seconds before logging. Mk3 stations use a continuously averaging software system. The wind readings are vector averaged, so direction is accounted for properly.

Calibration procedures and routines are detailed on the Engineering website.

3. Wind Sensor Specification

The following are additional specifications of the wind sensors used with Mk3 stations. A new sensor will be used with Mk5 stations. Wind sensors are mounted at a nominal 10 meters above water.

The R.M. Young sensor has the following characteristics:
* Wind Speed
Range: 0-60 m/s
Pitch: 29.4 cm air passes per rev.
Distance constant: 2.7 m for 63% recovery
* Wind Direction
Range: 360 deg, with 5 deg electrically open at north
Damping ratio: 0.25
Delay distance: 1.5 m for 50% recovery
Threshold: 1.0 m/s @ 10 deg.
Displacement: 1.5 m/s @ 5 deg. displacement
Damped w/length: 7.4 m
Undamped w/length: 7.2 m

4. Underwater Temperature Sensors

These sensors are interchangeable thermistors in Mk3 stations. They can be mounted a significant distance from the weather station, using a 2 wire connection. The basic accuracy is due to the use of interchangeable units. However improved accuracy is obtained by calibrating against a precision reference sensor in situ. These are in turn calibrated against a standard traceable to 0.04 degrees.

本数据集包含两套自动气象站（automatic weather station）采集的气象与海水温度数据，两套设备于不同时段系泊于汤斯维尔（Townsville）境内的克利夫兰湾（Cleveland Bay）不同点位。

首台气象站部署于10号航道浮标（Channel Marker #10）点位，坐标为南纬19°11.95′、东经146°51.89′。该站采集数据的时段为1990年7月3日至1992年10月21日，后续该设备被移除并重新系泊于克利夫兰角（Cape Cleveland）。

当前在用的气象站部署于S5号航道浮标点位，距离磁岛（Magnetic Island）上的布雷姆纳角（Bremner Point）约1公里，坐标为南纬19°09.35′、东经146°52.87′。该站自1999年12月7日起开始采集数据。

本数据集采集的数据用于支持澳大利亚海洋科学研究所（Australian Institute of Marine Science，AIMS）的科研工作。相关数据可应其他研究人员及公众的申请对外提供。

当前在用的气象站为澳大利亚坎贝尔科学（Campbell Scientific）系统。

采集的记录参数包括：海水温度（对应平均海平面1.9米与8.5米水深，Mean Sea Level，MSL）、大气压力、空气温度、太阳辐射（光合有效辐射，PAR）、真风向（标量平均）、真风速（每半小时前进行10分钟平均计算，并记录最大阵风风速）。

1. 运行机制与气象传感器

气象站每半小时采集并将数据存储于电子内存中。中心基站通过高频（HF）无线电或电话线定期轮询各远端气象站。数据以频移键控信号形式通过无线电传输，并封装为信息数据包。系统通过奇偶校验与校验和方法检测数据错误，基站识别到无效数据包后会请求重传，该机制可在通信质量不佳的情况下恢复绝大多数有效数据。远端气象站可本地存储数据达21天，系统集成了自动运行、远程控制、远程校时与内置诊断等功能。

传感器是气象站的核心组成部分，以下传感器的选型综合考量了成本、可靠性与精度指标。
* 风速风向传感器由R.M.Young公司生产，型号为05103。该传感器为螺旋桨式，具备高线性度、高互换性与低启动阈值的优势。风向测量结果以风的来向为准。
* 太阳辐射传感器采用Licor公司生产的水下量子传感器，以光合有效辐射（Photosynthetically Active Radiation，PAR）为单位测量光照强度，其光谱响应经过定义与加权校正。传感器滤光片老化导致的漂移是已知问题，同类产品均存在该情况。
* 温度传感器均采用Omega公司的可互换热敏电阻，该类传感器具备互换性强、精度高的特点，但可靠性存在一定问题，目前团队正考虑替换方案。
* 早期气象站采用经过改装的Aanderaa型气压传感器，Mk2型气象站搭载Weathertronics单元。目前所有气象站均为Mk3型，搭载的维萨拉（Vaisala）气压传感器具备更高的互换性与精度。

2. 系统精度

系统精度由以下各项误差之和计算得出：
* 校准误差
* 传感器互换误差
* 随时间产生的漂移误差
* 0~40℃环境温度范围带来的影响误差

以下为Mk3型气象站搭载传感器的规格参数。Mk5型气象站将采用全新的传感器套件，其选型部分基于维萨拉WXT510气象传感器。温度与风速传感器均支持互换，且无需单独校准，但部分传感器曾经过标准设备比对检测。
* 空气温度：可互换热敏电阻与电子组件的测量误差范围为±0.4℃，在空气中的稳定时间为30秒。低风速环境下，温度防护屏的抽吸效应会带来额外误差。
* 海水温度：可互换热敏电阻与电子组件的测量误差范围为±0.4℃，在水中的稳定时间为30分钟。通常会采用精度更高的现场校准（误差约±0.1℃），其校准结果可溯源至精度为0.04℃的标准传感器。
* 太阳辐射（PAR）：测量误差为读数的±5%，传感器初始年漂移量约为-4%。
* 大气压力：测量误差为±1百帕（hPa）。
* 风速：测量误差为读数的2% ±满量程的0.1%（FSD）。
* 风向：测量误差为读数的2% ±满量程的0.1%（FSD）。

太阳辐射与风速参数的电稳定时间为7秒，该参数符合抗混叠滤波器的设计要求。Mk1与Mk2型气象站在数据记录前，会在16秒内采集16个样本并取平均；Mk3型气象站则采用持续平均的软件系统。风速风向数据采用矢量平均法，可正确计算风向。校准流程与操作规范详见工程官网。

3. 风速传感器规格

以下为Mk3型气象站搭载的风速传感器的补充规格参数。Mk5型气象站将采用全新的风速传感器。风速传感器的标称安装高度为水面以上10米。
R.M.Young公司生产的该款传感器具备以下特性：
* 风速
测量范围：0~60 m/s
螺距：每旋转一圈气流通过29.4 cm
距离常数：63%响应恢复距离为2.7 m
* 风向
测量范围：360°，正北方向存在5°的电气开路区间
阻尼比：0.25
延迟距离：50%响应恢复距离为1.5 m
启动阈值：10°方位角下为1.0 m/s
位移响应：5°方位角偏移下响应为1.5 m/s
阻尼波长：7.4 m
无阻尼波长：7.2 m

4. 水下温度传感器

Mk3型气象站搭载的水下温度传感器为可互换热敏电阻，可通过两线制接线方式安装在距离气象站较远的点位。其基础精度源于可互换的传感器设计，但通过现场使用精密参考传感器进行校准，可进一步提升测量精度。该参考传感器本身亦可溯源至精度为0.04℃的标准传感器进行校准。