Vízügyi Közlemények, 1983 (65. évfolyam)
1. füzet - Hock Béla: Vízminőség-változási tendenciák vizsgálata
108 Hock Béla (соотношения (7) и (8)). Определение статистических параметров модели может быть выполнено при помощи множественной линейной регрессии. Имея в виду полученные результаты автор выводит оценку средней тенденции изменения качества воды для ряда с длиной К и числом элементов (соотношение (12)). Автор обращает внимание на исследование степени значимости выявленной тенденции. По практическому опыту предел значимости устанавливается цифрой 3%/год. Практическая оценка тенденций изменения качества родных ресурсов автором выполнена для 11 створов в бассейне Тисы (табл. I.) по недельным отборам проб в период 1970-79 гг. Оценка распространена на следующие компоненты качества воды : ВПК,, растворенный 0 2. Оз — • РОГ, сумма растворенных веществ. Результаты получены на основании рядов данных с числом элементов п = 40, полученных при учете средних значений по кварталам года, причем осреднение выполнено методом соотношений разбавления (соотношение (8)). Коэффициент множественной корреляции ( R) и результаты исследования значимости регрессии приводятся в габл. — III —IV. Степень усредненного изменения качества воды показана в табл. VI., значимость изменения — в табл. VI. Конечные результаты изображены на рис. I. # # * Trend analysis of changes in water quality by B. HOCK The main concern of water management in the industrialized countries all over the World is caused by the rapid and serious deterioration observable in the quality of water resources, which have already become one of the most important factors of production. The quality of surface waters in Hungary was rather poor a decade ago already. In the development of the present situation the neglect of providing adequate wastewater treatment and the widespread use of chemicals in agriculture have played the most important roles. The manner in which the quality deterioration of surface waters over an extended period of time is reflected by the measurement data, viz. the records of the sampling stations forming the standard network, as well as the methods by which the trend of quality changes can be determined are dealt with in the present paper. A review of the professional literature on the subject has revealed that the variations in the various components of water quality have been examined in terms of streamflow, water temperature or time as the independent variables. No attempt has thus far been undertaken to express the combined effect of these three variables. This may be attributed in part to the absence of scarcity of records of the required length and based on sufficiently frequent samplings. For solving the problem the author has introduced a general (linear) model Eq. (4) — and suggested subsequently a simplified version thereof — Eqs. (7)—(8) — for practical applications. A potential approach to determining the statistical parameters of the model consists of multiple linear regression analysis based on the method of least squares. The average deterioration of water quality is obtained therefrom for a record of lenght К and consisting of n elements — Eq. (12). Attention is called to the importance of examining the significance level of the change, for the limit of which 3%/year is suggested on the basis of practical experiences. The practical application of the trend analysis of quality changes is illustrated for eleven cross sections in the Tisza river system (Table 1), relying on samples taken each week over the period 1970 to 1979. All samples have been analysed for BOD s, dissolved 0 2, NH 4 + —N, N0 3~ —N, РОД and total dissolved solids. The data obtained have been averaged for quarter —year periods and the resulting time series of n = 40 elements ha been used in combination with the dilution — proportionate model Eq. (8) — to determine the results presented. The multiple correlation
