Homework Problem \#2 - Arctic Ice

• /1v1/08-LionGuo/02-Homework-Problem-2-Arctic-Iceproblem_2.ipynb

• /1v1/08-LionGuo/02-Homework-Problem-2-Arctic-Ice/day_2.ipynb

Hand-in format: IPython Notebook or python program. Submit via email.

As a reminder: please make sure your code is clean, documentated, and understandable. Make sure it runs without errors.

提交格式: IPython Notebook或python程序。通过电子邮件提交。

作为一个提醒:请确保您的代码是干净的、有文档记录的和可理解的。确保它没有错误地运行。

Background

The purpose of this problem is to become familiar with loading, manipulating, and analyzing image-like data, plotting it. We will use a dataset collected by the AMSR-E instrument Aquaopen in new window satellite.

这个问题的目的是熟悉加载、操作和分析类似图像的数据，并绘制它。我们将使用AMSR-E仪器Aquaopen in new window卫星收集的数据集。

The data consists of maps of the concentration of ice in the Arctic collected between 2006 and 2011. The data obtained from the amsr databaseopen in new window and converted into a single HDF5 file format.

这些数据由2006年至2011年收集的北极冰浓度地图组成。从amsr数据库open in new window获得的数据，并转换为单一的HDF5文件格式。

Part 1 - Examining a single map

第1部分-检查单个地图

Begin by examining the HDF5 file - you can use h5ls at the command line, or h5py inside the notebook.

从检查HDF5文件开始-你可以在命令行中使用' h5ls '，或者在笔记本中使用' h5py '。

If you don't remember how to open HDF5 files, and read datasets from HDF5 files, look at our Day 2 lecture.

如果你不记得如何打开 HDF5 文件，并从 HDF5 文件中读取数据集，请参考我们第2天的课程。

There are many datasets, each with a name of the format YYYYMMDD, giving the data. Each dataset is a single map (i.e. 2D array), where the values give the ice concentration (fraction, from 0.0 to 100.0) in that pixel of the map. Careful of NaN values!

有许多数据集，每个数据集的名称格式为“YYYYMMDD”，用于提供数据。每个数据集都是一个单一的地图(即2D数组)，其中的值给出了地图像素中的冰浓度(从0.0到100.0的分数)。小心NaN值!

Read one of the maps, and plot it with Matplotlib.

阅读其中一张地图，并使用Matplotlib绘制它。

Note: to get the correct orientation, you need the origin='lower' argument for imshow(). Include a colorbar. Remove the tick labels (0, 100, and so on, indicating pixel number) since they are not useful.

注意:为了获得正确的方向，你需要' origin='lower' '参数为' imshow() '。包括一个colorbar。删除标记(“0”、“100”等，表示像素数)，因为它们没有用。

答案

# -*- coding: utf-8 -*-
# @Time    : 2022/10/6 18:26
# @Author  : AI悦创
# @FileName: hw2.py
# @Software: PyCharm
# @Blog    ：https://bornforthis.cn/
import h5py


# f = h5py.File('p2_icedata_area.hdf5', 'r')
# # f = h5py.File('p2_icedata.hdf5', 'r')
# print(f)
# print(f.keys())
# dataset = f["pixel_areas"]
# print(dataset)
# print(dataset.shape)
# print(dataset.dtype)

def read_hdf5(path):
    f = h5py.File(path, 'r')
    keys = f.keys()
    for key in keys:
        # print(key)
        dataset = f[key]
        print(key, dataset.shape, dataset.dtype, sep="\t")


if __name__ == '__main__':
    read_hdf5("p2_icedata.hdf5")

import matplotlib.pyplot as plt
import h5py
def read_hdf5(path):
    year_lst = []
    dataset_lst = []
    f = h5py.File(path, 'r')
    keys = f.keys()
    for key in keys:
        # print(key)
        dataset = f[key]
#         print(key, dataset.shape, dataset.dtype, sep="\t")
        year_lst.append(key)
        dataset_lst.append(dataset.shape)
    # "g" 表示红色，marksize用来设置'D'菱形的大小
    plt.plot(year_lst, dataset_lst, "g", marker='D', markersize=5, label="year")
    # 绘制坐标轴标签
    plt.xlabel("年")
    plt.ylabel("数据")
    plt.title("HDF5")
    # 显示图例
    plt.legend(loc="lower right")
    # 调用 text()在图像上绘制注释文本
    # x1、y1表示文本所处坐标位置，ha参数控制水平对齐方式, va控制垂直对齐方式，str(y1)表示要绘制的文本
#     for x1, y1 in zip(year_lst, dataset_lst):
#         plt.text(x1, y1, str(y1), ha='center', va='bottom', fontsize=10)
    # 保存图片
    plt.savefig("1.jpg")
    plt.show()


read_hdf5("p2_icedata.hdf5")

Part 2 - Ice concentration versus time

第二部分-冰浓度与时间的关系

We want to make a plot of the ice concentration over time.

我们要画出冰浓度随时间变化的曲线。

First, write a loop to read all the datasets of the HDF5 file (e.g. into a dict).

首先，编写一个循环来读取HDF5文件的所有数据集(例如到字典中)。

Then, write an analysis function frac_pixels_above(dict,value) which, for each array in the input dict, computes the fraction of pixels above the input value. Use this to make a plot of the number of pixels with concentration above 50%, versus time.

然后，编写一个分析函数' frac_pixels_above(dict,value) '，该函数对于输入dict中的每个数组，计算输入' value '以上像素的百分比。使用此方法绘制浓度超过50%的像素数量与时间的关系图。

Note: to include "time" on the x-axis of a plot, you may want to write a helper function to convert the dict keys from their YYYYMMDD string format into a 3-tuple of (year, month, day) integer values.

注意:要在图形的x轴上包含“时间”，您可能需要编写一个helper函数来将字典键从它们的“YYYYMMDD”字符串格式转换为3元组(年、月、日)整数值。

This can then be converted into fractional years (e.g. 1 July 2012 is 2012.5). For simplicity you can assume each month has 30 days.

这可以转换为小数年(例如，2012年7月1日为“2012.5”)。为了简单起见，你可以假设每个月有30天。

Try experimenting with matplotlib set_major_formatter to get a good representation of dates in the tick labelsopen in new window.

尝试使用matplotlib ' set_major_formatter '来获得一个很好的在标记标签中日期的表示open in new window。

Describe what you see in the plot.

描述你在情节中看到了什么。

def read_hdf5(path):
    # year_lst = []
    dataset_lst = []
    f = h5py.File(path, 'r')
    keys = f.keys()
    for key in keys:
        # print(key)
        dataset = f[key]
        # print(key, dataset.shape, dataset.dtype, sep="\t")
        # year_lst.append(key)
        dataset_lst.append((key, dataset.shape))
    return dict(dataset_lst)

Part 3 - Physical units

第3部分-物理单元

To be more quantitative we will compute the actual surface area of Earth in $\rm{km}^2$ over which the ice concentration is above a given threshold.

为了更定量，我们将以$\rm{km}^2$为单位计算地球的实际表面积，在此范围内冰浓度超过给定的阈值。

However, these maps are projections of a spherical surface, so pixels have different areaopen in new window.

然而，这些地图是球面的投影，所以像素有不同的面积open in new window。

Every map uses the same projection, so the pixel areas in each are the same.

每个地图使用相同的投影，所以每个地图中的像素区域是相同的。

The areas (in $\rm{km}^2$) are available in the file named data/p2_icedata_area.hdf5. Inspect, then load, this datafile. Plot it (with colorbar and units).

区域(在$\rm{km}^2$中)在名为“data/p2_icedata_area.hdf5”的文件中可用。检查并加载这个数据文件。绘制它(用颜色条和单位)。

• p2_icedata_area.hdf5
• p2_icedata.hdf5

文件读取

欢迎关注我公众号：AI悦创，有更多更好玩的等你发现！

公众号：AI悦创【二维码】